GitHub/exynos8895/android_kernel_samsung_universal8895.git
8 years agoUPSTREAM: arm64: mm: add code to safely replace TTBR1_EL1
Mark Rutland [Mon, 25 Jan 2016 11:45:01 +0000 (11:45 +0000)]
UPSTREAM: arm64: mm: add code to safely replace TTBR1_EL1

If page tables are modified without suitable TLB maintenance, the ARM
architecture permits multiple TLB entries to be allocated for the same
VA. When this occurs, it is permitted that TLB conflict aborts are
raised in response to synchronous data/instruction accesses, and/or and
amalgamation of the TLB entries may be used as a result of a TLB lookup.

The presence of conflicting TLB entries may result in a variety of
behaviours detrimental to the system (e.g. erroneous physical addresses
may be used by I-cache fetches and/or page table walks). Some of these
cases may result in unexpected changes of hardware state, and/or result
in the (asynchronous) delivery of SError.

To avoid these issues, we must avoid situations where conflicting
entries may be allocated into TLBs. For user and module mappings we can
follow a strict break-before-make approach, but this cannot work for
modifications to the swapper page tables that cover the kernel text and
data.

Instead, this patch adds code which is intended to be executed from the
idmap, which can safely unmap the swapper page tables as it only
requires the idmap to be active. This enables us to uninstall the active
TTBR1_EL1 entry, invalidate TLBs, then install a new TTBR1_EL1 entry
without potentially unmapping code or data required for the sequence.
This avoids the risk of conflict, but requires that updates are staged
in a copy of the swapper page tables prior to being installed.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Cc: Laura Abbott <labbott@fedoraproject.org>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 50e1881ddde2a986c7d0d2150985239e5e3d7d96)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Ibd0a7c802a1abe0ba08a99819fd62ac646186005

8 years agoUPSTREAM: arm64: add function to install the idmap
Mark Rutland [Mon, 25 Jan 2016 11:45:00 +0000 (11:45 +0000)]
UPSTREAM: arm64: add function to install the idmap

In some cases (e.g. when making invasive changes to the kernel page
tables) we will need to execute code from the idmap.

Add a new helper which may be used to install the idmap, complementing
the existing cpu_uninstall_idmap.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Cc: Laura Abbott <labbott@fedoraproject.org>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 609116d202a8c5fd3fe393eb85373cbee906df68)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I1aac9e85e3d49add72c8aab7d80777a39f5fdd8e

8 years agoUPSTREAM: arm64: unmap idmap earlier
Mark Rutland [Mon, 25 Jan 2016 11:44:59 +0000 (11:44 +0000)]
UPSTREAM: arm64: unmap idmap earlier

During boot we leave the idmap in place until paging_init, as we
previously had to wait for the zero page to become allocated and
accessible.

Now that we have a statically-allocated zero page, we can uninstall the
idmap much earlier in the boot process, making it far easier to spot
accidental use of physical addresses. This also brings the cold boot
path in line with the secondary boot path.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Cc: Laura Abbott <labbott@fedoraproject.org>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 86ccce896cb0aa800a7a6dcd29b41ffc4eeb1a75)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I6375bd9855e45727790697875b7cd19f84a4dd7f

8 years agoUPSTREAM: arm64: unify idmap removal
Mark Rutland [Mon, 25 Jan 2016 11:44:58 +0000 (11:44 +0000)]
UPSTREAM: arm64: unify idmap removal

We currently open-code the removal of the idmap and restoration of the
current task's MMU state in a few places.

Before introducing yet more copies of this sequence, unify these to call
a new helper, cpu_uninstall_idmap.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Cc: Laura Abbott <labbott@fedoraproject.org>
Cc: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 9e8e865bbe294a69666a1996bda3e87825b258c0)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I6e9cb0253a1d2d63232f8fa0b3f39f8f6987b239

8 years agoUPSTREAM: arm64: mm: place empty_zero_page in bss
Mark Rutland [Mon, 25 Jan 2016 11:44:57 +0000 (11:44 +0000)]
UPSTREAM: arm64: mm: place empty_zero_page in bss

Currently the zero page is set up in paging_init, and thus we cannot use
the zero page earlier. We use the zero page as a reserved TTBR value
from which no TLB entries may be allocated (e.g. when uninstalling the
idmap). To enable such usage earlier (as may be required for invasive
changes to the kernel page tables), and to minimise the time that the
idmap is active, we need to be able to use the zero page before
paging_init.

This patch follows the example set by x86, by allocating the zero page
at compile time, in .bss. This means that the zero page itself is
available immediately upon entry to start_kernel (as we zero .bss before
this), and also means that the zero page takes up no space in the raw
Image binary. The associated struct page is allocated in bootmem_init,
and remains unavailable until this time.

Outside of arch code, the only users of empty_zero_page assume that the
empty_zero_page symbol refers to the zeroed memory itself, and that
ZERO_PAGE(x) must be used to acquire the associated struct page,
following the example of x86. This patch also brings arm64 inline with
these assumptions.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Cc: Laura Abbott <labbott@fedoraproject.org>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 5227cfa71f9e8574373f4d0e9e754942d76cdf67)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I43005dd8533d9968d2cef48bd2821d4507cae5ad

8 years agoUPSTREAM: arm64: mm: specialise pagetable allocators
Mark Rutland [Mon, 25 Jan 2016 11:44:56 +0000 (11:44 +0000)]
UPSTREAM: arm64: mm: specialise pagetable allocators

We pass a size parameter to early_alloc and late_alloc, but these are
only ever used to allocate single pages. In late_alloc we always
allocate a single page.

Both allocators provide us with zeroed pages (such that all entries are
invalid), but we have no barriers between allocating a page and adding
that page to existing (live) tables. A concurrent page table walk may
see stale data, leading to a number of issues.

This patch specialises the two allocators for page tables. The size
parameter is removed and the necessary dsb(ishst) is folded into each.
To make it clear that the functions are intended for use for page table
allocation, they are renamed to {early,late}_pgtable_alloc, with the
related function pointed renamed to pgtable_alloc.

As the dsb(ishst) is now in the allocator, the existing barrier for the
zero page is redundant and thus is removed. The previously missing
include of barrier.h is added.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Tested-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Tested-by: Jeremy Linton <jeremy.linton@arm.com>
Cc: Laura Abbott <labbott@fedoraproject.org>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 21ab99c289d350f4ae454bc069870009db6df20e)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Ib88fafc146506943122aa1ca6ee5a6c331ddb26c

8 years agoUPSTREAM: asm-generic: Fix local variable shadow in __set_fixmap_offset
Mark Rutland [Mon, 25 Jan 2016 11:44:55 +0000 (11:44 +0000)]
UPSTREAM: asm-generic: Fix local variable shadow in __set_fixmap_offset

Currently __set_fixmap_offset is a macro function which has a local
variable called 'addr'. If a caller passes a 'phys' parameter which is
derived from a variable also called 'addr', the local variable will
shadow this, and the compiler will complain about the use of an
uninitialized variable. To avoid the issue with namespace clashes,
'addr' is prefixed with a liberal sprinkling of underscores.

Turning __set_fixmap_offset into a static inline breaks the build for
several architectures. Fixing this properly requires updates to a number
of architectures to make them agree on the prototype of __set_fixmap (it
could be done as a subsequent patch series).

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Arnd Bergmann <arnd@arndb.de>
[catalin.marinas@arm.com: squashed the original function patch and macro fixup]
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 3694bd76781b76c4f8d2ecd85018feeb1609f0e5)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Iec27cb36dfca39e333de9f7318e76da0670d0156

8 years agoBACKPORT: Eliminate the .eh_frame sections from the aarch64 vmlinux and kernel modules
William Cohen [Fri, 22 Jan 2016 03:56:26 +0000 (22:56 -0500)]
BACKPORT: Eliminate the .eh_frame sections from the aarch64 vmlinux and kernel modules

By default the aarch64 gcc generates .eh_frame sections.  Unlike
.debug_frame sections, the .eh_frame sections are loaded into memory
when the associated code is loaded.  On an example kernel being built
with this default the .eh_frame section in vmlinux used an extra 1.7MB
of memory.  The x86 disables the creation of the .eh_frame section.
The aarch64 should probably do the same to save some memory.

Signed-off-by: William Cohen <wcohen@redhat.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 728dabd6d1751cf5e0f8e0535891393da62396e9)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I697baae2209d6d11f2cc447459d935f7200eb7b1

8 years agoUPSTREAM: arm64: Fix an enum typo in mm/dump.c
Masanari Iida [Sun, 24 Jan 2016 06:24:12 +0000 (15:24 +0900)]
UPSTREAM: arm64: Fix an enum typo in mm/dump.c

This patch fixes a typo in mm/dump.c:
"MODUELS_END_NR" should be "MODULES_END_NR".

Signed-off-by: Masanari Iida <standby24x7@gmail.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit b3122023df935cf14bf951da98ca598d71b9f826)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I5e73d18fd70f20f200a974f5f2ba22cb7ae64952

8 years agoUPSTREAM: arm64: kasan: ensure that the KASAN zero page is mapped read-only
Ard Biesheuvel [Mon, 11 Jan 2016 13:50:21 +0000 (14:50 +0100)]
UPSTREAM: arm64: kasan: ensure that the KASAN zero page is mapped read-only

When switching from the early KASAN shadow region, which maps the
entire shadow space read-write, to the permanent KASAN shadow region,
which uses a zero page to shadow regions that are not subject to
instrumentation, the lowest level table kasan_zero_pte[] may be
reused unmodified, which means that the mappings of the zero page
that it contains will still be read-write.

So update it explicitly to map the zero page read only when we
activate the permanent mapping.

Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 7b1af9795773d745c2a8c7d4ca5f2936e8b6adfb)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I4745dc91236c2612ad3e13be1cc176ffd923f7da

8 years agoUPSTREAM: arch/arm/include/asm/pgtable-3level.h: add pmd_mkclean for THP
Minchan Kim [Sat, 16 Jan 2016 00:55:33 +0000 (16:55 -0800)]
UPSTREAM: arch/arm/include/asm/pgtable-3level.h: add pmd_mkclean for THP

MADV_FREE needs pmd_dirty and pmd_mkclean for detecting recent overwrite
of the contents since MADV_FREE syscall is called for THP page.

This patch adds pmd_mkclean for THP page MADV_FREE support.

Signed-off-by: Minchan Kim <minchan@kernel.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: Russell King <rmk@arm.linux.org.uk>
Cc: "James E.J. Bottomley" <jejb@parisc-linux.org>
Cc: "Kirill A. Shutemov" <kirill@shutemov.name>
Cc: Shaohua Li <shli@kernel.org>
Cc: <yalin.wang2010@gmail.com>
Cc: Andrea Arcangeli <aarcange@redhat.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Chen Gang <gang.chen.5i5j@gmail.com>
Cc: Chris Zankel <chris@zankel.net>
Cc: Daniel Micay <danielmicay@gmail.com>
Cc: Darrick J. Wong <darrick.wong@oracle.com>
Cc: David S. Miller <davem@davemloft.net>
Cc: Helge Deller <deller@gmx.de>
Cc: Hugh Dickins <hughd@google.com>
Cc: Ivan Kokshaysky <ink@jurassic.park.msu.ru>
Cc: Jason Evans <je@fb.com>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com>
Cc: Kirill A. Shutemov <kirill@shutemov.name>
Cc: Matt Turner <mattst88@gmail.com>
Cc: Max Filippov <jcmvbkbc@gmail.com>
Cc: Mel Gorman <mgorman@suse.de>
Cc: Michael Kerrisk <mtk.manpages@gmail.com>
Cc: Michal Hocko <mhocko@suse.com>
Cc: Mika Penttil <mika.penttila@nextfour.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Richard Henderson <rth@twiddle.net>
Cc: Rik van Riel <riel@redhat.com>
Cc: Roland Dreier <roland@kernel.org>
Cc: Shaohua Li <shli@kernel.org>
Cc: Wu Fengguang <fengguang.wu@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 44842045e4baaf406db2954dd2e07152fa61528d)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I59d53667aa8c40dea4f18fc58acc7d27f4a85a04

8 years agoUPSTREAM: arm64: hide __efistub_ aliases from kallsyms
Ard Biesheuvel [Fri, 15 Jan 2016 12:28:57 +0000 (13:28 +0100)]
UPSTREAM: arm64: hide __efistub_ aliases from kallsyms

Commit e8f3010f7326 ("arm64/efi: isolate EFI stub from the kernel
proper") isolated the EFI stub code from the kernel proper by prefixing
all of its symbols with __efistub_, and selectively allowing access to
core kernel symbols from the stub by emitting __efistub_ aliases for
functions and variables that the stub can access legally.

As an unintended side effect, these aliases are emitted into the
kallsyms symbol table, which means they may turn up in backtraces,
e.g.,

  ...
  PC is at __efistub_memset+0x108/0x200
  LR is at fixup_init+0x3c/0x48
  ...
  [<ffffff8008328608>] __efistub_memset+0x108/0x200
  [<ffffff8008094dcc>] free_initmem+0x2c/0x40
  [<ffffff8008645198>] kernel_init+0x20/0xe0
  [<ffffff8008085cd0>] ret_from_fork+0x10/0x40

The backtrace in question has nothing to do with the EFI stub, but
simply returns one of the several aliases of memset() that have been
recorded in the kallsyms table. This is undesirable, since it may
suggest to people who are not aware of this that the issue they are
seeing is somehow EFI related.

So hide the __efistub_ aliases from kallsyms, by emitting them as
absolute linker symbols explicitly. The distinction between those
and section relative symbols is completely irrelevant to these
definitions, and to the final link we are performing when these
definitions are being taken into account (the distinction is only
relevant to symbols defined inside a section definition when performing
a partial link), and so the resulting values are identical to the
original ones. Since absolute symbols are ignored by kallsyms, this
will result in these values to be omitted from its symbol table.

After this patch, the backtrace generated from the same address looks
like this:
  ...
  PC is at __memset+0x108/0x200
  LR is at fixup_init+0x3c/0x48
  ...
  [<ffffff8008328608>] __memset+0x108/0x200
  [<ffffff8008094dcc>] free_initmem+0x2c/0x40
  [<ffffff8008645198>] kernel_init+0x20/0xe0
  [<ffffff8008085cd0>] ret_from_fork+0x10/0x40

Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 75feee3d9d51775072d3a04f47d4a439a4c4590e)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Ieeb8c576e31f0dd0b7f82982ffa6362864eed311

8 years agoUPSTREAM: arm64: head.S: use memset to clear BSS
Mark Rutland [Wed, 6 Jan 2016 11:05:27 +0000 (11:05 +0000)]
UPSTREAM: arm64: head.S: use memset to clear BSS

Currently we use an open-coded memzero to clear the BSS. As it is a
trivial implementation, it is sub-optimal.

Our optimised memset doesn't use the stack, is position-independent, and
for the memzero case can use of DC ZVA to clear large blocks
efficiently. In __mmap_switched the MMU is on and there are no live
caller-saved registers, so we can safely call an uninstrumented memset.

This patch changes __mmap_switched to use memset when clearing the BSS.
We use the __pi_memset alias so as to avoid any instrumentation in all
kernel configurations.

Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Marc Zyngier <marc.zyngier@arm.com>
Reviewed-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 2a803c4db615d85126c5c7afd5849a3cfde71422)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I3dc7050fe5566f2126cbea9abfa6063c8e6b029a

8 years agoUPSTREAM: efi: stub: define DISABLE_BRANCH_PROFILING for all architectures
Ard Biesheuvel [Wed, 23 Dec 2015 09:29:28 +0000 (10:29 +0100)]
UPSTREAM: efi: stub: define DISABLE_BRANCH_PROFILING for all architectures

This moves the DISABLE_BRANCH_PROFILING define from the x86 specific
to the general CFLAGS definition for the stub. This fixes build errors
when building for arm64 with CONFIG_PROFILE_ALL_BRANCHES_ENABLED.

Reviewed-by: Matt Fleming <matt@codeblueprint.co.uk>
Reported-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit b523e185bba36164ca48a190f5468c140d815414)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Id5a2a3986adc60ddf3d247c038667ce9719f3ee0

8 years agoUPSTREAM: arm64: entry: remove pointless SPSR mode check
Mark Rutland [Tue, 5 Jan 2016 17:33:34 +0000 (17:33 +0000)]
UPSTREAM: arm64: entry: remove pointless SPSR mode check

In work_pending, we may skip work if the stacked SPSR value represents
anything other than an EL0 context. We then immediately invoke the
kernel_exit 0 macro as part of ret_to_user, assuming a return to EL0.
This is somewhat confusing.

We use work_pending as part of the ret_to_user/ret_fast_syscall state
machine. We only use ret_fast_syscall in the return from an SVC issued
from EL0. We use ret_to_user for return from EL0 exception handlers and
also for return from ret_from_fork in the case the task was not a kernel
thread (i.e. it is a user task).

Thus in all cases the stacked SPSR value must represent an EL0 context,
and the check is redundant. This patch removes it, along with the now
unused no_work_pending label.

Cc: Chris Metcalf <cmetcalf@ezchip.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit ee03353bc04f8e460cc4e3da80d9721d9ecb89f1)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I2738149342ef469e8cde7c5f8f7c65daab93fb2b

8 years agoUPSTREAM: arm64: mm: move pgd_cache initialisation to pgtable_cache_init
Will Deacon [Tue, 5 Jan 2016 15:36:59 +0000 (15:36 +0000)]
UPSTREAM: arm64: mm: move pgd_cache initialisation to pgtable_cache_init

Initialising the suppport for EFI runtime services requires us to
allocate a pgd off the back of an early_initcall. On systems where the
PGD_SIZE is smaller than PAGE_SIZE (e.g. 64k pages and 48-bit VA), the
pgd_cache isn't initialised at this stage, and we panic with a NULL
dereference during boot:

  Unable to handle kernel NULL pointer dereference at virtual address 00000000

  __create_mapping.isra.5+0x84/0x350
  create_pgd_mapping+0x20/0x28
  efi_create_mapping+0x5c/0x6c
  arm_enable_runtime_services+0x154/0x1e4
  do_one_initcall+0x8c/0x190
  kernel_init_freeable+0x84/0x1ec
  kernel_init+0x10/0xe0
  ret_from_fork+0x10/0x50

This patch fixes the problem by initialising the pgd_cache earlier, in
the pgtable_cache_init callback, which sounds suspiciously like what it
was intended for.

Reported-by: Dennis Chen <dennis.chen@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 39b5be9b4233a9f212b98242bddf008f379b5122)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I124f03d19299be93124af35641294bf73c13bb22

8 years agoUPSTREAM: arm64: traps: address fallout from printk -> pr_* conversion
Will Deacon [Mon, 21 Dec 2015 16:44:27 +0000 (16:44 +0000)]
UPSTREAM: arm64: traps: address fallout from printk -> pr_* conversion

Commit ac7b406c1a9d ("arm64: Use pr_* instead of printk") was a fairly
mindless s/printk/pr_*/ change driven by a complaint from checkpatch.

As is usual with such changes, this has led to some odd behaviour on
arm64:

  * syslog now picks up the "pr_emerg" line from dump_backtrace, but not
    the actual trace, which leads to a bunch of "kernel:Call trace:"
    lines in the log

  * __{pte,pmd,pgd}_error print at KERN_CRIT, as opposed to KERN_ERR
    which is used by other architectures.

This patch restores the original printk behaviour for dump_backtrace
and downgrade the pgtable error macros to KERN_ERR.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit c9cd0ed925c0b927283d4739bfe689eb9d1e9dfd)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Iaf028e5368df4623f7257ef432a7f9da86261609

8 years agoUPSTREAM: arm64: ftrace: fix a stack tracer's output under function graph tracer
AKASHI Takahiro [Tue, 15 Dec 2015 08:33:41 +0000 (17:33 +0900)]
UPSTREAM: arm64: ftrace: fix a stack tracer's output under function graph tracer

Function graph tracer modifies a return address (LR) in a stack frame
to hook a function return. This will result in many useless entries
(return_to_handler) showing up in
 a) a stack tracer's output
 b) perf call graph (with perf record -g)
 c) dump_backtrace (at panic et al.)

For example, in case of a),
  $ echo function_graph > /sys/kernel/debug/tracing/current_tracer
  $ echo 1 > /proc/sys/kernel/stack_trace_enabled
  $ cat /sys/kernel/debug/tracing/stack_trace
        Depth    Size   Location    (54 entries)
        -----    ----   --------
  0)     4504      16   gic_raise_softirq+0x28/0x150
  1)     4488      80   smp_cross_call+0x38/0xb8
  2)     4408      48   return_to_handler+0x0/0x40
  3)     4360      32   return_to_handler+0x0/0x40
  ...

In case of b),
  $ echo function_graph > /sys/kernel/debug/tracing/current_tracer
  $ perf record -e mem:XXX:x -ag -- sleep 10
  $ perf report
                  ...
                  |          |          |--0.22%-- 0x550f8
                  |          |          |          0x10888
                  |          |          |          el0_svc_naked
                  |          |          |          sys_openat
                  |          |          |          return_to_handler
                  |          |          |          return_to_handler
                  ...

In case of c),
  $ echo function_graph > /sys/kernel/debug/tracing/current_tracer
  $ echo c > /proc/sysrq-trigger
  ...
  Call trace:
  [<ffffffc00044d3ac>] sysrq_handle_crash+0x24/0x30
  [<ffffffc000092250>] return_to_handler+0x0/0x40
  [<ffffffc000092250>] return_to_handler+0x0/0x40
  ...

This patch replaces such entries with real addresses preserved in
current->ret_stack[] at unwind_frame(). This way, we can cover all
the cases.

Reviewed-by: Jungseok Lee <jungseoklee85@gmail.com>
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
[will: fixed minor context changes conflicting with irq stack bits]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 20380bb390a443b2c5c8800cec59743faf8151b4)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I6360182f8d04fdd2e31c0cb6054aefa2adb216e7

8 years agoUPSTREAM: arm64: pass a task parameter to unwind_frame()
AKASHI Takahiro [Tue, 15 Dec 2015 08:33:40 +0000 (17:33 +0900)]
UPSTREAM: arm64: pass a task parameter to unwind_frame()

Function graph tracer modifies a return address (LR) in a stack frame
to hook a function's return. This will result in many useless entries
(return_to_handler) showing up in a call stack list.
We will fix this problem in a later patch ("arm64: ftrace: fix a stack
tracer's output under function graph tracer"). But since real return
addresses are saved in ret_stack[] array in struct task_struct,
unwind functions need to be notified of, in addition to a stack pointer
address, which task is being traced in order to find out real return
addresses.

This patch extends unwind functions' interfaces by adding an extra
argument of a pointer to task_struct.

Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit fe13f95b720075327a761fe6ddb45b0c90cab504)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I92a9a07468c182d5abbacaa73a90984ab11ad535

8 years agoUPSTREAM: arm64: ftrace: modify a stack frame in a safe way
AKASHI Takahiro [Tue, 15 Dec 2015 08:33:39 +0000 (17:33 +0900)]
UPSTREAM: arm64: ftrace: modify a stack frame in a safe way

Function graph tracer modifies a return address (LR) in a stack frame by
calling ftrace_prepare_return() in a traced function's function prologue.
The current code does this modification before preserving an original
address at ftrace_push_return_trace() and there is always a small window
of inconsistency when an interrupt occurs.

This doesn't matter, as far as an interrupt stack is introduced, because
stack tracer won't be invoked in an interrupt context. But it would be
better to proactively minimize such a window by moving the LR modification
after ftrace_push_return_trace().

Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 79fdee9b6355c9720f14717e1ad66af51bb331b5)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Ief0a136aa01348b4d0d3f447544f21fd77835c67

8 years agoUPSTREAM: arm64: remove irq_count and do_softirq_own_stack()
James Morse [Fri, 18 Dec 2015 16:01:47 +0000 (16:01 +0000)]
UPSTREAM: arm64: remove irq_count and do_softirq_own_stack()

sysrq_handle_reboot() re-enables interrupts while on the irq stack. The
irq_stack implementation wrongly assumed this would only ever happen
via the softirq path, allowing it to update irq_count late, in
do_softirq_own_stack().

This means if an irq occurs in sysrq_handle_reboot(), during
emergency_restart() the stack will be corrupted, as irq_count wasn't
updated.

Lose the optimisation, and instead of moving the adding/subtracting of
irq_count into irq_stack_entry/irq_stack_exit, remove it, and compare
sp_el0 (struct thread_info) with sp & ~(THREAD_SIZE - 1). This tells us
if we are on a task stack, if so, we can safely switch to the irq stack.
Finally, remove do_softirq_own_stack(), we don't need it anymore.

Reported-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: James Morse <james.morse@arm.com>
[will: use get_thread_info macro]
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit d224a69e3d80fe08f285d1f41d21b590bae4fa9f)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I1f613279bf875443b10d65b1cd1ed4a6abfcb605

8 years agoUPSTREAM: arm64: hugetlb: add support for PTE contiguous bit
David Woods [Thu, 17 Dec 2015 19:31:26 +0000 (14:31 -0500)]
UPSTREAM: arm64: hugetlb: add support for PTE contiguous bit

The arm64 MMU supports a Contiguous bit which is a hint that the TTE
is one of a set of contiguous entries which can be cached in a single
TLB entry.  Supporting this bit adds new intermediate huge page sizes.

The set of huge page sizes available depends on the base page size.
Without using contiguous pages the huge page sizes are as follows.

 4KB:   2MB  1GB
64KB: 512MB

With a 4KB granule, the contiguous bit groups together sets of 16 pages
and with a 64KB granule it groups sets of 32 pages.  This enables two new
huge page sizes in each case, so that the full set of available sizes
is as follows.

 4KB:  64KB   2MB  32MB  1GB
64KB:   2MB 512MB  16GB

If a 16KB granule is used then the contiguous bit groups 128 pages
at the PTE level and 32 pages at the PMD level.

If the base page size is set to 64KB then 2MB pages are enabled by
default.  It is possible in the future to make 2MB the default huge
page size for both 4KB and 64KB granules.

Reviewed-by: Chris Metcalf <cmetcalf@ezchip.com>
Reviewed-by: Steve Capper <steve.capper@linaro.org>
Signed-off-by: David Woods <dwoods@ezchip.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 66b3923a1a0f77a563b43f43f6ad091354abbfe9)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I5e99c5165bc5eb966adf4d4523632fd9eedd9602

8 years agoBACKPORT: arm64: Use PoU cache instr for I/D coherency
Ashok Kumar [Thu, 17 Dec 2015 09:38:32 +0000 (01:38 -0800)]
BACKPORT: arm64: Use PoU cache instr for I/D coherency

In systems with three levels of cache(PoU at L1 and PoC at L3),
PoC cache flush instructions flushes L2 and L3 caches which could affect
performance.
For cache flushes for I and D coherency, PoU should suffice.
So changing all I and D coherency related cache flushes to PoU.

Introduced a new __clean_dcache_area_pou API for dcache flush till PoU
and provided a common macro for __flush_dcache_area and
__clean_dcache_area_pou.

Also, now in __sync_icache_dcache, icache invalidation for non-aliasing
VIPT icache is done only for that particular page instead of the earlier
__flush_icache_all.

Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Ashok Kumar <ashoks@broadcom.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 0a28714c53fd4f7aea709be7577dfbe0095c8c3e)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I64f065140d5e8783e91ed53ae9c7a2e33a3e515a

8 years agoBACKPORT: arm64: kernel: fix architected PMU registers unconditional access
Lorenzo Pieralisi [Wed, 13 Jan 2016 14:50:03 +0000 (14:50 +0000)]
BACKPORT: arm64: kernel: fix architected PMU registers unconditional access

The Performance Monitors extension is an optional feature of the
AArch64 architecture, therefore, in order to access Performance
Monitors registers safely, the kernel should detect the architected
PMU unit presence through the ID_AA64DFR0_EL1 register PMUVer field
before accessing them.

This patch implements a guard by reading the ID_AA64DFR0_EL1 register
PMUVer field to detect the architected PMU presence and prevent accessing
PMU system registers if the Performance Monitors extension is not
implemented in the core.

Cc: Peter Maydell <peter.maydell@linaro.org>
Cc: Mark Rutland <mark.rutland@arm.com>
Cc: <stable@vger.kernel.org>
Fixes: 60792ad349f3 ("arm64: kernel: enforce pmuserenr_el0 initialization and restore")
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reported-by: Guenter Roeck <linux@roeck-us.net>
Tested-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit f436b2ac90a095746beb6729b8ee8ed87c9eaede)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Ie442b9feba5d143bd6b6c82d70190fc8bc95298d

8 years agoUPSTREAM: arm64: Defer dcache flush in __cpu_copy_user_page
Ashok Kumar [Thu, 17 Dec 2015 09:38:31 +0000 (01:38 -0800)]
UPSTREAM: arm64: Defer dcache flush in __cpu_copy_user_page

Defer dcache flushing to __sync_icache_dcache by calling
flush_dcache_page which clears PG_dcache_clean flag.

Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Ashok Kumar <ashoks@broadcom.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit e6b1185f77351aa154e63bd54b05d07ff99d4ffa)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I2e02ce2f43f68287337bed30e3c3455c0eee4164

8 years agoUPSTREAM: arm64: reduce stack use in irq_handler
James Morse [Tue, 15 Dec 2015 11:21:25 +0000 (11:21 +0000)]
UPSTREAM: arm64: reduce stack use in irq_handler

The code for switching to irq_stack stores three pieces of information on
the stack, fp+lr, as a fake stack frame (that lets us walk back onto the
interrupted tasks stack frame), and the address of the struct pt_regs that
contains the register values from kernel entry. (which dump_backtrace()
will print in any stack trace).

To reduce this, we store fp, and the pointer to the struct pt_regs.
unwind_frame() can recognise this as the irq_stack dummy frame, (as it only
appears at the top of the irq_stack), and use the struct pt_regs values
to find the missing interrupted link-register.

Suggested-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 971c67ce37cfeeaf560e792a2c3bc21d8b67163a)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I84cbb04857a441083d331e875c3e228d24ec2276

8 years agoUPSTREAM: arm64: Documentation: add list of software workarounds for errata
Will Deacon [Tue, 17 Nov 2015 14:45:47 +0000 (14:45 +0000)]
UPSTREAM: arm64: Documentation: add list of software workarounds for errata

It's not immediately obvious which hardware errata are worked around in
the Linux kernel for an arbitrary kernel tree, so add a file to keep
track of what we're working around.

Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 9cb9c9e5ba8453537e8e645318edf231fe54eaf9)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I139972ff6e10e12c4b4f27cd047f55b3dd8b4118

8 years agoUPSTREAM: arm64: mm: place __cpu_setup in .text
Mark Rutland [Fri, 11 Dec 2015 11:04:31 +0000 (11:04 +0000)]
UPSTREAM: arm64: mm: place __cpu_setup in .text

We drop __cpu_setup in .text.init, which ends up being part of .text.
The .text.init section was a legacy section name which has been unused
elsewhere for a long time.

The ".text.init" name is misleading if read as a synonym for
".init.text". Any CPU may execute __cpu_setup before turning the MMU on,
so it should simply live in .text.

Remove the pointless section assignment. This will leave __cpu_setup in
the .text section.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit f00083cae331e5d3eecade6b4fdc35d0825e73ef)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I2e9b154cd6de92662c70c2b957479448252c661a

8 years agoUPSTREAM: arm64: cmpxchg: Don't incldue linux/mmdebug.h
Mark Brown [Thu, 10 Dec 2015 16:54:32 +0000 (16:54 +0000)]
UPSTREAM: arm64: cmpxchg: Don't incldue linux/mmdebug.h

The arm64 asm/cmpxchg.h includes linux/mmdebug.h but doesn't so far as I
can tell actually use anything from it.  Removing the inclusion reduces
spurious header dependency rebuilds and also avoids issues with
recursive inclusions of headers causing build breaks due to attempts to
use things before they are defined if linux/mmdebug.h starts pulling in
more low level headers.

Such errors have happened in -next recently, for example:

In file included from include/linux/completion.h:11:0,
                 from include/linux/rcupdate.h:43,
                 from include/linux/tracepoint.h:19,
                 from include/linux/mmdebug.h:6,
                 from ./arch/arm64/include/asm/cmpxchg.h:22,
                 from ./arch/arm64/include/asm/atomic.h:41,
                 from include/linux/atomic.h:4,
                 from include/linux/spinlock.h:406,
                 from include/linux/seqlock.h:35,
                 from include/linux/time.h:5,
                 from include/uapi/linux/timex.h:56,
                 from include/linux/timex.h:56,
                 from include/linux/sched.h:19,
                 from arch/arm64/kernel/asm-offsets.c:21:
include/linux/wait.h: In function 'wait_on_atomic_t':
include/linux/wait.h:1218:2: error: implicit declaration of function 'atomic_read' [-Werror=implicit-function-declaration]
 if (atomic_read(val) == 0)

Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 4a6ccf30263f4e265c0f171561bf4c40bed5f273)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Idf44176dad0abc11e67b4e416b02a3fba14f3f1b

8 years agoUPSTREAM: arm64: mm: fold alternatives into .init
Mark Rutland [Wed, 9 Dec 2015 12:44:38 +0000 (12:44 +0000)]
UPSTREAM: arm64: mm: fold alternatives into .init

Currently we treat the alternatives separately from other data that's
only used during initialisation, using separate .altinstructions and
.altinstr_replacement linker sections. These are freed for general
allocation separately from .init*. This is problematic as:

* We do not remove execute permissions, as we do for .init, leaving the
  memory executable.

* We pad between them, making the kernel Image bianry up to PAGE_SIZE
  bytes larger than necessary.

This patch moves the two sections into the contiguous region used for
.init*. This saves some memory, ensures that we remove execute
permissions, and allows us to remove some code made redundant by this
reorganisation.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Andre Przywara <andre.przywara@arm.com>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Jeremy Linton <jeremy.linton@arm.com>
Cc: Laura Abbott <labbott@fedoraproject.org>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 9aa4ec1571da62366cfddc20f3b923609604fe63)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I3fee118ead5c73ade50ea10d436881c9424a549c

8 years agoBACKPORT: arm64: Remove redundant padding from linker script
Mark Rutland [Wed, 9 Dec 2015 12:44:37 +0000 (12:44 +0000)]
BACKPORT: arm64: Remove redundant padding from linker script

Currently we place an ALIGN_DEBUG_RO between text and data for the .text
and .init sections, and depending on configuration each of these may
result in up to SECTION_SIZE bytes worth of padding (for
DEBUG_RODATA_ALIGN).

We make no distinction between the text and data in each of these
sections at any point when creating the initial page tables in head.S.
We also make no distinction when modifying the tables; __map_memblock,
fixup_executable, mark_rodata_ro, and fixup_init only work at section
granularity. Thus this padding is unnecessary.

For the spit between init text and data we impose a minimum alignment of
16 bytes, but this is also unnecessary. The init data is output
immediately after the padding before any symbols are defined, so this is
not required to keep a symbol for linker a section array correctly
associated with the data. Any objects within the section will be given
at least their usual alignment regardless.

This patch removes the redundant padding.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Jeremy Linton <jeremy.linton@arm.com>
Cc: Laura Abbott <labbott@fedoraproject.org>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit 5b28cd9d084eca8ddc46270d2720305bfd40e348)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I5540ba1f4d90e2ae8fafa22e98c389bc4e975ac7

8 years agoUPSTREAM: arm64: mm: remove pointless PAGE_MASKing
Mark Rutland [Wed, 9 Dec 2015 12:44:36 +0000 (12:44 +0000)]
UPSTREAM: arm64: mm: remove pointless PAGE_MASKing

As pgd_offset{,_k} shift the input address by PGDIR_SHIFT, the sub-page
bits will always be shifted out. There is no need to apply PAGE_MASK
before this.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Cc: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Jeremy Linton <jeremy.linton@arm.com>
Cc: Laura Abbott <labbott@fedoraproject.org>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: rework-pagetable

(cherry picked from commit e2c30ee320eb96304896c7ab84499e5bc5e5fb6e)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I86d3438aecf7295d5895e1430c1e19fbc82c9719

8 years agonet: inet: diag: expose the socket mark to privileged processes.
Lorenzo Colitti [Wed, 7 Sep 2016 15:42:25 +0000 (00:42 +0900)]
net: inet: diag: expose the socket mark to privileged processes.

This adds the capability for a process that has CAP_NET_ADMIN on
a socket to see the socket mark in socket dumps.

Commit a52e95abf772 ("net: diag: allow socket bytecode filters to
match socket marks") recently gave privileged processes the
ability to filter socket dumps based on mark. This patch is
complementary: it ensures that the mark is also passed to
userspace in the socket's netlink attributes.  It is useful for
tools like ss which display information about sockets.

[backport of net-next d545caca827b65aab557a9e9dcdcf1e5a3823c2d]

Change-Id: I33336ed9c3ee3fb78fe05c4c47b7fd18c6e33ef1
Tested: https://android-review.googlesource.com/270210
Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: diag: make udp_diag_destroy work for mapped addresses.
Lorenzo Colitti [Wed, 7 Sep 2016 04:38:35 +0000 (13:38 +0900)]
net: diag: make udp_diag_destroy work for mapped addresses.

udp_diag_destroy does look up the IPv4 UDP hashtable for mapped
addresses, but it gets the IPv4 address to look up from the
beginning of the IPv6 address instead of the end.

[cherry-pick of net-next f95bf346226b9b79352e05508beececc807cc37a]

Change-Id: Ia369482c4645bcade320b2c33a763f1ce4378ff1
Tested: https://android-review.googlesource.com/269874
Fixes: 5d77dca82839 ("net: diag: support SOCK_DESTROY for UDP sockets")
Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
Acked-by: Eric Dumazet <edumazet@google.com>
Acked-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: diag: support SOCK_DESTROY for UDP sockets
David Ahern [Wed, 24 Aug 2016 04:06:33 +0000 (21:06 -0700)]
net: diag: support SOCK_DESTROY for UDP sockets

This implements SOCK_DESTROY for UDP sockets similar to what was done
for TCP with commit c1e64e298b8ca ("net: diag: Support destroying TCP
sockets.") A process with a UDP socket targeted for destroy is awakened
and recvmsg fails with ECONNABORTED.

[cherry-pick of 5d77dca82839ef016a93ad7acd7058b14d967752]

Change-Id: I4b4862548e6e3c05dde27781e7daa0b18b93bd81
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: diag: allow socket bytecode filters to match socket marks
Lorenzo Colitti [Wed, 24 Aug 2016 06:46:26 +0000 (15:46 +0900)]
net: diag: allow socket bytecode filters to match socket marks

This allows a privileged process to filter by socket mark when
dumping sockets via INET_DIAG_BY_FAMILY. This is useful on
systems that use mark-based routing such as Android.

The ability to filter socket marks requires CAP_NET_ADMIN, which
is consistent with other privileged operations allowed by the
SOCK_DIAG interface such as the ability to destroy sockets and
the ability to inspect BPF filters attached to packet sockets.

[cherry-pick of a52e95abf772b43c9226e9a72d3c1353903ba96f]

Change-Id: I8b90b814264d9808bda050cdba8f104943bdb9a8
Tested: https://android-review.googlesource.com/261350
Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
Acked-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: diag: slightly refactor the inet_diag_bc_audit error checks.
Lorenzo Colitti [Wed, 24 Aug 2016 06:46:25 +0000 (15:46 +0900)]
net: diag: slightly refactor the inet_diag_bc_audit error checks.

This simplifies the code a bit and also allows inet_diag_bc_audit
to send to userspace an error that isn't EINVAL.

[cherry-pick of net-next 627cc4add53c0470bfd118002669205d222d3a54]

Change-Id: Iee3d2bbb19f3110d71f0698ffb293f9bdffc8ef1
Signed-off-by: Lorenzo Colitti <lorenzo@google.com>
Acked-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agonet: diag: Add support to filter on device index
David Ahern [Fri, 24 Jun 2016 01:42:51 +0000 (18:42 -0700)]
net: diag: Add support to filter on device index

Add support to inet_diag facility to filter sockets based on device
index. If an interface index is in the filter only sockets bound
to that index (sk_bound_dev_if) are returned.

[cherry-pick of net-next 637c841dd7a5f9bd97b75cbe90b526fa1a52e530]

Change-Id: I6b6bcdcf15d3142003f1ee53b4d82f2fabbb8250
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
8 years agoUPSTREAM: arm64: don't call C code with el0's fp register
James Morse [Thu, 10 Dec 2015 10:22:41 +0000 (10:22 +0000)]
UPSTREAM: arm64: don't call C code with el0's fp register

On entry from el0, we save all the registers on the kernel stack, and
restore them before returning. x29 remains unchanged when we call out
to C code, which will store x29 as the frame-pointer on the stack.

Instead, write 0 into x29 after entry from el0, to avoid any risk of
tracing into user space.

Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: per-cpu-irq-stack

(cherry picked from commit 49003a8d6b35e128ef5e51433e60e783a46fbe5f)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Ifae7003018e4088d5de038cef25fa210211a75b6

8 years agoUPSTREAM: arm64: when walking onto the task stack, check sp & fp are in current-...
James Morse [Thu, 10 Dec 2015 10:22:40 +0000 (10:22 +0000)]
UPSTREAM: arm64: when walking onto the task stack, check sp & fp are in current->stack

When unwind_frame() reaches the bottom of the irq_stack, the last fp
points to the original task stack. unwind_frame() uses
IRQ_STACK_TO_TASK_STACK() to find the sp value. If either values is
wrong, we may end up walking a corrupt stack.

Check these values are sane by testing if they are both on the stack
pointed to by current->stack.

Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: per-cpu-irq-stack

(cherry picked from commit 1ffe199b1c9b72a8e752a9ae2a7af10128ab2ca1)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I2e5bf1ce899a1018f1c5b8ccb4f7c816d61bba21

8 years agoUPSTREAM: arm64: Add this_cpu_ptr() assembler macro for use in entry.S
James Morse [Thu, 10 Dec 2015 10:22:39 +0000 (10:22 +0000)]
UPSTREAM: arm64: Add this_cpu_ptr() assembler macro for use in entry.S

irq_stack is a per_cpu variable, that needs to be access from entry.S.
Use an assembler macro instead of the unreadable details.

Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: per-cpu-irq-stack

(cherry picked from commit aa4d5d3cbc258c355151a3903211b27359390ec5)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I2ccc11f442c0303c62e1c3e0a05a088958c922b8

8 years agoUPSTREAM: arm64: irq: fix walking from irq stack to task stack
Will Deacon [Wed, 9 Dec 2015 13:58:42 +0000 (13:58 +0000)]
UPSTREAM: arm64: irq: fix walking from irq stack to task stack

Running with CONFIG_DEBUG_SPINLOCK=y can trigger a BUG with the new IRQ
stack code:

  BUG: spinlock lockup suspected on CPU#1

This is due to the IRQ_STACK_TO_TASK_STACK macro incorrectly retrieving
the task stack pointer stashed at the top of the IRQ stack.

Sayeth James:

| Yup, this is what is happening. Its an off-by-one due to broken
| thinking about how the stack works. My broken thinking was:
|
| >   top ------------
| >       | dummy_lr | <- irq_stack_ptr
| >       ------------
| >       |   x29    |
| >       ------------
| >       |   x19    | <- irq_stack_ptr - 0x10
| >       ------------
| >       |   xzr    |
| >       ------------
|
| But the stack-pointer is decreased before use. So it actually looks
| like this:
|
| >       ------------
| >       |          |  <- irq_stack_ptr
| >   top ------------
| >       | dummy_lr |
| >       ------------
| >       |   x29    | <- irq_stack_ptr - 0x10
| >       ------------
| >       |   x19    |
| >       ------------
| >       |   xzr    | <- irq_stack_ptr - 0x20
| >       ------------
|
| The value being used as the original stack is x29, which in all the
| tests is sp but without the current frames data, hence there are no
| missing frames in the output.
|
| Jungseok Lee picked it up with a 32bit user space because aarch32
| can't use x29, so it remains 0 forever. The fix he posted is correct.

This patch fixes the macro and adds some of this wisdom to a comment,
so that the layout of the IRQ stack is well understood.

Cc: James Morse <james.morse@arm.com>
Reported-by: Jungseok Lee <jungseoklee85@gmail.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: per-cpu-irq-stack

(cherry picked from commit 7596abf2e5661d52c4f414f37addeed54e098880)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Ic65c0d0d90a30a5caf8b3791d1e856400bd2b5f5

8 years agoUPSTREAM: arm64: Add do_softirq_own_stack() and enable irq_stacks
James Morse [Fri, 4 Dec 2015 11:02:27 +0000 (11:02 +0000)]
UPSTREAM: arm64: Add do_softirq_own_stack() and enable irq_stacks

entry.S is modified to switch to the per_cpu irq_stack during el{0,1}_irq.
irq_count is used to detect recursive interrupts on the irq_stack, it is
updated late by do_softirq_own_stack(), when called on the irq_stack, before
__do_softirq() re-enables interrupts to process softirqs.

do_softirq_own_stack() is added by this patch, but does not yet switch
stack.

This patch adds the dummy stack frame and data needed by the previous
stack tracing patches.

Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: per-cpu-irq-stack

(cherry picked from commit 8e23dacd12a48e58125b84c817da50850b73280a)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I9f79437af3da0398cb12e7afd1aa9f473f59b2e6

8 years agoUPSTREAM: arm64: Modify stack trace and dump for use with irq_stack
AKASHI Takahiro [Fri, 4 Dec 2015 11:02:26 +0000 (11:02 +0000)]
UPSTREAM: arm64: Modify stack trace and dump for use with irq_stack

This patch allows unwind_frame() to traverse from interrupt stack to task
stack correctly. It requires data from a dummy stack frame, created
during irq_stack_entry(), added by a later patch.

A similar approach is taken to modify dump_backtrace(), which expects to
find struct pt_regs underneath any call to functions marked __exception.
When on an irq_stack, the struct pt_regs is stored on the old task stack,
the location of which is stored in the dummy stack frame.

Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
[james.morse: merged two patches, reworked for per_cpu irq_stacks, and
 no alignment guarantees, added irq_stack definitions]
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: per-cpu-irq-stack

(cherry picked from commit 132cd887b5c54758d04bf25c52fa48f45e843a30)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I60b29291620a71ab7b6564730299d29f41ceb199

8 years agoUPSTREAM: arm64: Store struct thread_info in sp_el0
Jungseok Lee [Fri, 4 Dec 2015 11:02:25 +0000 (11:02 +0000)]
UPSTREAM: arm64: Store struct thread_info in sp_el0

There is need for figuring out how to manage struct thread_info data when
IRQ stack is introduced. struct thread_info information should be copied
to IRQ stack under the current thread_info calculation logic whenever
context switching is invoked. This is too expensive to keep supporting
the approach.

Instead, this patch pays attention to sp_el0 which is an unused scratch
register in EL1 context. sp_el0 utilization not only simplifies the
management, but also prevents text section size from being increased
largely due to static allocated IRQ stack as removing masking operation
using THREAD_SIZE in many places.

Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Jungseok Lee <jungseoklee85@gmail.com>
Signed-off-by: James Morse <james.morse@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: per-cpu-irq-stack

(cherry picked from commit 6cdf9c7ca687e01840d0215437620a20263012fc)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I53c9f44a0772b8649f302a65a7a6519d8eebcb91

8 years agoUPSTREAM: arm64: Add trace_hardirqs_off annotation in ret_to_user
Catalin Marinas [Fri, 4 Dec 2015 12:42:29 +0000 (12:42 +0000)]
UPSTREAM: arm64: Add trace_hardirqs_off annotation in ret_to_user

When a kernel is built with CONFIG_TRACE_IRQFLAGS the following warning
is produced when entering userspace for the first time:

  WARNING: at /work/Linux/linux-2.6-aarch64/kernel/locking/lockdep.c:3519
  Modules linked in:
  CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc3+ #639
  Hardware name: Juno (DT)
  task: ffffffc9768a0000 ti: ffffffc9768a8000 task.ti: ffffffc9768a8000
  PC is at check_flags.part.22+0x19c/0x1a8
  LR is at check_flags.part.22+0x19c/0x1a8
  pc : [<ffffffc0000fba6c>] lr : [<ffffffc0000fba6c>] pstate: 600001c5
  sp : ffffffc9768abe10
  x29: ffffffc9768abe10 x28: ffffffc9768a8000
  x27: 0000000000000000 x26: 0000000000000001
  x25: 00000000000000a6 x24: ffffffc00064be6c
  x23: ffffffc0009f249e x22: ffffffc9768a0000
  x21: ffffffc97fea5480 x20: 00000000000001c0
  x19: ffffffc00169a000 x18: 0000005558cc7b58
  x17: 0000007fb78e3180 x16: 0000005558d2e238
  x15: ffffffffffffffff x14: 0ffffffffffffffd
  x13: 0000000000000008 x12: 0101010101010101
  x11: 7f7f7f7f7f7f7f7f x10: fefefefefefeff63
  x9 : 7f7f7f7f7f7f7f7f x8 : 6e655f7371726964
  x7 : 0000000000000001 x6 : ffffffc0001079c4
  x5 : 0000000000000000 x4 : 0000000000000001
  x3 : ffffffc001698438 x2 : 0000000000000000
  x1 : ffffffc9768a0000 x0 : 000000000000002e
  Call trace:
  [<ffffffc0000fba6c>] check_flags.part.22+0x19c/0x1a8
  [<ffffffc0000fc440>] lock_is_held+0x80/0x98
  [<ffffffc00064bafc>] __schedule+0x404/0x730
  [<ffffffc00064be6c>] schedule+0x44/0xb8
  [<ffffffc000085bb0>] ret_to_user+0x0/0x24
  possible reason: unannotated irqs-off.
  irq event stamp: 502169
  hardirqs last  enabled at (502169): [<ffffffc000085a98>] el0_irq_naked+0x1c/0x24
  hardirqs last disabled at (502167): [<ffffffc0000bb3bc>] __do_softirq+0x17c/0x298
  softirqs last  enabled at (502168): [<ffffffc0000bb43c>] __do_softirq+0x1fc/0x298
  softirqs last disabled at (502143): [<ffffffc0000bb830>] irq_exit+0xa0/0xf0

This happens because we disable interrupts in ret_to_user before calling
schedule() in work_resched. This patch adds the necessary
trace_hardirqs_off annotation.

Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Reported-by: Mark Rutland <mark.rutland@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029

(cherry picked from commit db3899a6477a4dccd26cbfb7f408b6be2cc068e0)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I868c6c64c548f12156467959e2b8df09caae282f

8 years agoUPSTREAM: arm64: ftrace: fix the comments for ftrace_modify_code
Li Bin [Fri, 4 Dec 2015 03:38:40 +0000 (11:38 +0800)]
UPSTREAM: arm64: ftrace: fix the comments for ftrace_modify_code

There is no need to worry about module and __init text disappearing
case, because that ftrace has a module notifier that is called when
a module is being unloaded and before the text goes away and this
code grabs the ftrace_lock mutex and removes the module functions
from the ftrace list, such that it will no longer do any
modifications to that module's text, the update to make functions
be traced or not is done under the ftrace_lock mutex as well.
And by now, __init section codes should not been modified
by ftrace, because it is black listed in recordmcount.c and
ignored by ftrace.

Suggested-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Li Bin <huawei.libin@huawei.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: arm64-ftrace

(cherry picked from commit 004ab584e028093996cf5b8e220b8bc50c5111cf)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I69df7eddbf9e17031920b950312399dc4d36c09e

8 years agoUPSTREAM: arm64: ftrace: stop using kstop_machine to enable/disable tracing
Li Bin [Fri, 4 Dec 2015 03:38:39 +0000 (11:38 +0800)]
UPSTREAM: arm64: ftrace: stop using kstop_machine to enable/disable tracing

For ftrace on arm64, kstop_machine which is hugely disruptive
to a running system is not needed to convert nops to ftrace calls
or back, because that to be modified instrucions, that NOP, B or BL,
are all safe instructions which called "concurrent modification
and execution of instructions", that can be executed by one
thread of execution as they are being modified by another thread
of execution without requiring explicit synchronization.

Signed-off-by: Li Bin <huawei.libin@huawei.com>
Reviewed-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: arm64-ftrace

(cherry picked from commit 81a6a146e88eca5d6726569779778d61489d85aa)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I54e2c0d49bd68f9547bd9f0da8b7520e02bb0714

8 years agoUPSTREAM: arm64: spinlock: serialise spin_unlock_wait against concurrent lockers
Will Deacon [Thu, 19 Nov 2015 17:48:31 +0000 (17:48 +0000)]
UPSTREAM: arm64: spinlock: serialise spin_unlock_wait against concurrent lockers

Boqun Feng reported a rather nasty ordering issue with spin_unlock_wait
on architectures implementing spin_lock with LL/SC sequences and acquire
semantics:

 | CPU 1                   CPU 2                     CPU 3
 | ==================      ====================      ==============
 |                                                   spin_unlock(&lock);
 |                         spin_lock(&lock):
 |                           r1 = *lock; // r1 == 0;
 |                         o = READ_ONCE(object); // reordered here
 | object = NULL;
 | smp_mb();
 | spin_unlock_wait(&lock);
 |                           *lock = 1;
 | smp_mb();
 | o->dead = true;
 |                         if (o) // true
 |                           BUG_ON(o->dead); // true!!

The crux of the problem is that spin_unlock_wait(&lock) can return on
CPU 1 whilst CPU 2 is in the process of taking the lock. This can be
resolved by upgrading spin_unlock_wait to a LOCK operation, forcing it
to serialise against a concurrent locker and giving it acquire semantics
in the process (although it is not at all clear whether this is needed -
different callers seem to assume different things about the barrier
semantics and architectures are similarly disjoint in their
implementations of the macro).

This patch implements spin_unlock_wait using an LL/SC sequence with
acquire semantics on arm64. For v8.1 systems with the LSE atomics, the
exclusive writeback is omitted, since the spin_lock operation is
indivisible and no intermediate state can be observed.

Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029

(cherry picked from commit d86b8da04dfa4771a68bdbad6c424d40f22f0d14)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I27d88458c99dfd475d0326bd1d408ec3e575a7dd

8 years agoUPSTREAM: arm64: enable HAVE_IRQ_TIME_ACCOUNTING
Will Deacon [Mon, 23 Nov 2015 15:12:59 +0000 (15:12 +0000)]
UPSTREAM: arm64: enable HAVE_IRQ_TIME_ACCOUNTING

arm64 relies on the arm_arch_timer for sched_clock, so we can select
HAVE_IRQ_TIME_ACCOUNTING and have the core sched-clock code enable the
feature at runtime based on the rate.

Reported-by: Mario Smarduch <m.smarduch@samsung.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029

(cherry picked from commit 24da208db32ee1e4757ceaba898c47add8e5361e)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I849e010459dbbde0ac0d44d665dadcea9f8bf12d

8 years agoUPSTREAM: arm64: fix COMPAT_SHMLBA definition for large pages
Yury Norov [Wed, 2 Dec 2015 14:00:10 +0000 (14:00 +0000)]
UPSTREAM: arm64: fix COMPAT_SHMLBA definition for large pages

ARM glibc uses (4 * __getpagesize()) for SHMLBA, which is correct for
4KB pages and works fine for 64KB pages, but the kernel uses a hardcoded
16KB that is too small for 64KB page based kernels. This changes the
definition to what user space sees when using 64KB pages.

Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Yury Norov <ynorov@caviumnetworks.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029

(cherry picked from commit b9b7aebb42d1b1392f3111de61136bb6cf3aae3f)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I698814d005a28c7fe3963cded9756f88660d4aa0

8 years agoUPSTREAM: arm64: add __init/__initdata section marker to some functions/variables
Jisheng Zhang [Fri, 20 Nov 2015 09:59:10 +0000 (17:59 +0800)]
UPSTREAM: arm64: add __init/__initdata section marker to some functions/variables

These functions/variables are not needed after booting, so mark them
as __init or __initdata.

Signed-off-by: Jisheng Zhang <jszhang@marvell.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029

(cherry picked from commit a7c61a3452d39078919f0e1f493ff966fb64f0db)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I50a3362e186750e139d2440d2c1e1d49ace896e1

8 years agoUPSTREAM: usb: gadget: f_fs: Fix use-after-free
Lars-Peter Clausen [Thu, 14 Apr 2016 15:01:17 +0000 (17:01 +0200)]
UPSTREAM: usb: gadget: f_fs: Fix use-after-free

(cherry picked from commit 38740a5b87d53ceb89eb2c970150f6e94e00373a)

When using asynchronous read or write operations on the USB endpoints the
issuer of the IO request is notified by calling the ki_complete() callback
of the submitted kiocb when the URB has been completed.

Calling this ki_complete() callback will free kiocb. Make sure that the
structure is no longer accessed beyond that point, otherwise undefined
behaviour might occur.

Fixes: 2e4c7553cd6f ("usb: gadget: f_fs: add aio support")
Cc: <stable@vger.kernel.org> # v3.15+
Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Change-Id: I3c7b643f6440c4fb6160a57c1058523030b46a6c
Bug: 30950866

8 years agoUPSTREAM: brcmfmac: avoid potential stack overflow in brcmf_cfg80211_start_ap()
Arend van Spriel [Fri, 2 Sep 2016 08:37:24 +0000 (09:37 +0100)]
UPSTREAM: brcmfmac: avoid potential stack overflow in brcmf_cfg80211_start_ap()

commit ded89912156b1a47d940a0c954c43afbabd0c42c upstream

User-space can choose to omit NL80211_ATTR_SSID and only provide raw
IE TLV data. When doing so it can provide SSID IE with length exceeding
the allowed size. The driver further processes this IE copying it
into a local variable without checking the length. Hence stack can be
corrupted and used as exploit.

Cc: stable@vger.kernel.org # v4.4, v4.1
Reported-by: Daxing Guo <freener.gdx@gmail.com>
Reviewed-by: Hante Meuleman <hante.meuleman@broadcom.com>
Reviewed-by: Pieter-Paul Giesberts <pieter-paul.giesberts@broadcom.com>
Reviewed-by: Franky Lin <franky.lin@broadcom.com>
Signed-off-by: Arend van Spriel <arend.vanspriel@broadcom.com>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
8 years agoUPSTREAM: arm64: pgtable: implement pte_accessible()
Will Deacon [Fri, 30 Oct 2015 18:56:19 +0000 (18:56 +0000)]
UPSTREAM: arm64: pgtable: implement pte_accessible()

This patch implements the pte_accessible() macro, which can be used to
test whether or not a given pte is a candidate for allocation in the
TLB.

Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029

(cherry picked from commit 76c714be0e5e60c935a53b31be58939510ba1d0f)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I249e2d15665870149dd17d1cdb3850008f5a56fd

8 years agoUPSTREAM: arm64: mm: allow sections for unaligned bases
Mark Rutland [Mon, 23 Nov 2015 13:26:20 +0000 (13:26 +0000)]
UPSTREAM: arm64: mm: allow sections for unaligned bases

Callees of __create_mapping may decide to create section mappings if
sufficient low bits of the physical and virtual addresses they were
passed are zero. While __create_mapping rounds the virtual base address
down, it does not similarly round the physical base address down, and
hence non-zero bits in the physical address can prevent use of a section
mapping, even where a whole next-level table would be used instead.

Round down the physical base address in __create_mapping to enable all
callees to always create section mappings when such a mapping is
possible.

Cc: Laura Abbott <labbott@fedoraproject.org>
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Steve Capper <steve.capper@linaro.org>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: __create_mapping-fixes

(cherry picked from commit 9c4e08a3022b6df90d31ef4007291faabfce5431)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: Ic447350efdba3cd9f9e101c72183e04e39dd28d2

8 years agoUPSTREAM: arm64: mm: detect bad __create_mapping uses
Mark Rutland [Mon, 23 Nov 2015 13:26:19 +0000 (13:26 +0000)]
UPSTREAM: arm64: mm: detect bad __create_mapping uses

If a caller of __create_mapping provides a PA and VA which have
different sub-page offsets, it is not clear which offset they expect to
apply to the mapping, and is indicative of a bad caller.

In some cases, the region we wish to map may validly have a sub-page
offset in the physical and virtual addresses. For example, EFI runtime
regions have 4K granularity, yet may be mapped by a 64K page kernel. So
long as the physical and virtual offsets are the same, the region will
be mapped at the expected VAs.

Disallow calls with differing sub-page offsets, and WARN when they are
encountered, so that we can detect and fix such cases.

Cc: Laura Abbott <labbott@fedoraproject.org>
Acked-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Steve Capper <steve.capper@linaro.org>
Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Bug: 30369029
Patchset: __create_mapping-fixes

(cherry picked from commit cc5d2b3b95cdbb3fed4e38e667d17b9ac7250f7a)
Signed-off-by: Jeff Vander Stoep <jeffv@google.com>
Change-Id: I114a1265b10ff76daff385728d2125e618c313a1

8 years agoUPSTREAM: x86: fix SMAP in 32-bit environments
Linus Torvalds [Tue, 23 Feb 2016 22:58:52 +0000 (14:58 -0800)]
UPSTREAM: x86: fix SMAP in 32-bit environments

(cherry picked from commit de9e478b9d49f3a0214310d921450cf5bb4a21e6)

In commit 11f1a4b9755f ("x86: reorganize SMAP handling in user space
accesses") I changed how the stac/clac instructions were generated
around the user space accesses, which then made it possible to do
batched accesses efficiently for user string copies etc.

However, in doing so, I completely spaced out, and didn't even think
about the 32-bit case.  And nobody really even seemed to notice, because
SMAP doesn't even exist until modern Skylake processors, and you'd have
to be crazy to run 32-bit kernels on a modern CPU.

Which brings us to Andy Lutomirski.

He actually tested the 32-bit kernel on new hardware, and noticed that
it doesn't work.  My bad.  The trivial fix is to add the required
uaccess begin/end markers around the raw accesses in <asm/uaccess_32.h>.

I feel a bit bad about this patch, just because that header file really
should be cleaned up to avoid all the duplicated code in it, and this
commit just expands on the problem.  But this just fixes the bug without
any bigger cleanup surgery.

Reported-and-tested-by: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Change-Id: Ic044ebfe658a13179984111d062ca3a0b1404110
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
8 years agoUPSTREAM: audit: fix a double fetch in audit_log_single_execve_arg()
Paul Moore [Tue, 19 Jul 2016 21:42:57 +0000 (17:42 -0400)]
UPSTREAM: audit: fix a double fetch in audit_log_single_execve_arg()

(cherry picked from commit 43761473c254b45883a64441dd0bc85a42f3645c)

There is a double fetch problem in audit_log_single_execve_arg()
where we first check the execve(2) argumnets for any "bad" characters
which would require hex encoding and then re-fetch the arguments for
logging in the audit record[1].  Of course this leaves a window of
opportunity for an unsavory application to munge with the data.

This patch reworks things by only fetching the argument data once[2]
into a buffer where it is scanned and logged into the audit
records(s).  In addition to fixing the double fetch, this patch
improves on the original code in a few other ways: better handling
of large arguments which require encoding, stricter record length
checking, and some performance improvements (completely unverified,
but we got rid of some strlen() calls, that's got to be a good
thing).

As part of the development of this patch, I've also created a basic
regression test for the audit-testsuite, the test can be tracked on
GitHub at the following link:

 * https://github.com/linux-audit/audit-testsuite/issues/25

[1] If you pay careful attention, there is actually a triple fetch
problem due to a strnlen_user() call at the top of the function.

[2] This is a tiny white lie, we do make a call to strnlen_user()
prior to fetching the argument data.  I don't like it, but due to the
way the audit record is structured we really have no choice unless we
copy the entire argument at once (which would require a rather
wasteful allocation).  The good news is that with this patch the
kernel no longer relies on this strnlen_user() value for anything
beyond recording it in the log, we also update it with a trustworthy
value whenever possible.

Reported-by: Pengfei Wang <wpengfeinudt@gmail.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Paul Moore <paul@paul-moore.com>
Signed-off-by: Sasha Levin <alexander.levin@verizon.com>
Change-Id: I10e979e94605e3cf8d461e3e521f8f9837228aa5
Bug: 30956807

8 years agoUPSTREAM: ARM: 8494/1: mm: Enable PXN when running non-LPAE kernel on LPAE processor
Jungseung Lee [Mon, 28 Dec 2015 20:47:00 +0000 (04:47 +0800)]
UPSTREAM: ARM: 8494/1: mm: Enable PXN when running non-LPAE kernel on LPAE processor

The VMSA field of MMFR0 (bottom 4 bits) is incremented for each
added feature.  PXN is supported if the value is >= 4 and LPAE
is supported if it is >= 5.

In case a kernel with CONFIG_ARM_LPAE disabled is used on a
processor that supports LPAE, we can still use PXN in short
descriptors.  So check for >= 4 not == 4.

Signed-off-by: Jungseung Lee <js07.lee@samsung.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Ben Hutchings <ben@decadent.org.uk>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
8 years agoFIXUP: sched/tune: update accouting before CPU capacity
Patrick Bellasi [Wed, 24 Aug 2016 10:27:27 +0000 (11:27 +0100)]
FIXUP: sched/tune: update accouting before CPU capacity

The SchedTune tasks accounting is used to identify how many tasks are in
a boostgroup and thus to bias the selection of an OPP based on the
maximum boost value of the active boostgroups.
The current implementation however update the accounting after CPU
capacity has been update. This has two effects:
a) when we enqueue a boosted task, we do not immediately boost its CPU
b) when we dequeue a boosted task, we can keep a CPU boosted even if not
   required

This patch change the order of the SchedTune accounting and SchedFreq
updated to ensure to have always an updated representation of which
boosted tasks are runnable on a CPU before updating its capacity.

Reported-by: Leo Yan <leo.yan@linaro.org>
Signed-off-by: Patrick Bellasi <patrick.bellasi@arm.com>
8 years agoFIXUP: sched/tune: add fixes missing from a previous patch
Patrick Bellasi [Wed, 24 Aug 2016 10:02:29 +0000 (11:02 +0100)]
FIXUP: sched/tune: add fixes missing from a previous patch

The previous patch:

  e7ce26f - FIXUP: sched/tune: fix accounting for runnable tasks

squashed together patches of a series to fix SchedTune's accounting
issues. However, in the consolidation and cleanup of the series to merge
in the Android Common Kernel, we somehow missed a couple of important
changes:
1) the schedtune_exit function is not more required, because e7ce26f
   fixes accounting of exiting tasks in a different way
2) the schedtune_initialized flag was not set at the end of
   scheddtune_init_cgroup() thus failing to enabled SchedTune at boot.

This patch thus is to be considered an integration of e7ce26f.

Signed-off-by: Patrick Bellasi <patrick.bellasi@arm.com>
[jstultz: Cherry-picked from android-3.18. It should be noted that
some of this patch was already applied in the 4.4 patches (schedtune_exit
doesn't exist for example), but this patch just ensures things are totally
synced up]
Signed-off-by: John Stultz <john.stultz@linaro.org>
8 years agoarm: Fix #if/#ifdef typo in topology.c
Jon Medhurst [Thu, 2 Jun 2016 12:18:08 +0000 (12:18 +0000)]
arm: Fix #if/#ifdef typo in topology.c

Probably a typo in arch/arm/kernel/topology.c

This patch fixes the warning...

arch/arm/kernel/topology.c: In function 'scale_cpu_capacity':
arch/arm/kernel/topology.c:47:5: warning: "CONFIG_CPU_FREQ" is not defined [-Wundef]

Fixes: Change-Id: If5e9e0ba8ff5a5d3236b373dbce8c72ea71b5e18
       ("arm: Enable max freq invariant scheduler load-tracking and capacity support")
Signed-off-by: Jon Medhurst <tixy@linaro.org>
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
[jstultz: Cherry-picked from android-3.18]
Signed-off-by: John Stultz <john.stultz@linaro.org>
8 years agoarm: Fix build error "conflicting types for 'scale_cpu_capacity'"
Steve Muckle [Thu, 5 May 2016 01:56:45 +0000 (18:56 -0700)]
arm: Fix build error "conflicting types for 'scale_cpu_capacity'"

Commit "arm: Update arch_scale_cpu_capacity() to reflect change to
define" introduced a dependency on struct sched_domain in
arch/arm/include/asm/topologoy.h, but that structure is only currently
defined if CONFIG_CPU_FREQ is enabled, which causes
include/linux/cpufreq.h to get pulled in which defines it.

Include <linux/cpufreq.h> regardless of CONFIG_CPU_FREQ so struct
sched_domain is always defined.

Fixes: Change-Id: I372bd5e4c1e203428d72b18c8a806b06f3567ef6
       ("arm: Update arch_scale_cpu_capacity() to reflect change to define")
Signed-off-by: Steve Muckle <smuckle@linaro.org>
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
[jstultz: Cherry-picked from android-3.18]
Signed-off-by: John Stultz <john.stultz@linaro.org>
8 years agosched/walt: use do_div instead of division operator
Amit Pundir [Wed, 24 Aug 2016 06:22:17 +0000 (11:52 +0530)]
sched/walt: use do_div instead of division operator

Use do_div() instead of "/" operator to fix undefined references to
"__aeabi_uldivmod" build error for ARCH=arm.

Also in TP_fast_assign(), along with do_div() usage,  replace "," with
";" which would have resulted in a syntax error (!), because
'#define TP_fast_assign(args...) args' would have stripped off the ","
and left white space between these two assignments after CPP phase.

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
[jstultz: Cherry-picked from common/android-3.18]
Signed-off-by: John Stultz <john.stultz@linaro.org>
8 years agoDEBUG: cpufreq: fix cpu_capacity tracing build for non-smp systems
Amit Pundir [Mon, 29 Aug 2016 14:18:17 +0000 (19:48 +0530)]
DEBUG: cpufreq: fix cpu_capacity tracing build for non-smp systems

cpu curr capacity can only be traced for SMP systems. Non-SMP builds
will fail with:

drivers/cpufreq/cpufreq.c: In function â€˜cpufreq_freq_transition_begin’:
drivers/cpufreq/cpufreq.c:438:22: error: implicit declaration of function â€˜capacity_curr_of’ [-Werror=implicit-function-declaration]
   trace_cpu_capacity(capacity_curr_of(cpu), cpu);
                      ^

Fixes: Change-Id: Icd0930d11068fcb7d2b6a9a48e7ed974904e1081
       ("DEBUG: sched,cpufreq: add cpu_capacity change tracepoint")
Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
[jstultz: Cherry-picked from common/android-3.18]
Signed-off-by: John Stultz <john.stultz@linaro.org>
8 years agoUPSTREAM: ALSA: usb-audio: Fix double-free in error paths after snd_usb_add_audio_str...
Vladis Dronov [Thu, 31 Mar 2016 16:05:43 +0000 (12:05 -0400)]
UPSTREAM: ALSA: usb-audio: Fix double-free in error paths after snd_usb_add_audio_stream() call

(cherry picked from commit 836b34a935abc91e13e63053d0a83b24dfb5ea78)

create_fixed_stream_quirk(), snd_usb_parse_audio_interface() and
create_uaxx_quirk() functions allocate the audioformat object by themselves
and free it upon error before returning. However, once the object is linked
to a stream, it's freed again in snd_usb_audio_pcm_free(), thus it'll be
double-freed, eventually resulting in a memory corruption.

This patch fixes these failures in the error paths by unlinking the audioformat
object before freeing it.

Based on a patch by Takashi Iwai <tiwai@suse.de>

[Note for stable backports:
 this patch requires the commit 902eb7fd1e4a ('ALSA: usb-audio: Minor
 code cleanup in create_fixed_stream_quirk()')]

Bugzilla: https://bugzilla.redhat.com/show_bug.cgi?id=1283358
Reported-by: Ralf Spenneberg <ralf@spenneberg.net>
Cc: <stable@vger.kernel.org> # see the note above
Signed-off-by: Vladis Dronov <vdronov@redhat.com>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Change-Id: I7073a17d8c99886d2f6ed7981892712ba7dd5873
Bug: 30952477

8 years agoBACKPORT: ALSA: usb-audio: Minor code cleanup in create_fixed_stream_quirk()
Takashi Iwai [Tue, 15 Mar 2016 11:14:49 +0000 (12:14 +0100)]
BACKPORT: ALSA: usb-audio: Minor code cleanup in create_fixed_stream_quirk()

(cherry picked from commit 902eb7fd1e4af3ac69b9b30f8373f118c92b9729)

Just a minor code cleanup: unify the error paths.

Signed-off-by: Takashi Iwai <tiwai@suse.de>
Change-Id: I8253a86235df2ac1258153c9e128fa158527567f
Bug: 30952477

8 years agosched/walt: include missing header for arm_timer_read_counter()
Amit Pundir [Thu, 25 Aug 2016 05:36:37 +0000 (11:06 +0530)]
sched/walt: include missing header for arm_timer_read_counter()

Include clocksource/arm_arch_timer.h to fix implicit function
declaration of â€˜arch_timer_read_counter’ build error for ARCH=arm.

Signed-off-by: Amit Pundir <amit.pundir@linaro.org>
[jstultz: Cherry-picked from common/android-3.18]
Signed-off-by: John Stultz <john.stultz@linaro.org>
8 years agoMerge remote-tracking branch 'linaro-ext/EAS/v4.4-easv5.2+aosp-changes' into android-4.4
Dmitry Shmidt [Fri, 9 Sep 2016 00:02:15 +0000 (17:02 -0700)]
Merge remote-tracking branch 'linaro-ext/EAS/v4.4-easv5.2+aosp-changes' into android-4.4

Change-Id: Ic24b43ee867bc4f70b31bedaad734717b64b86a1

8 years agocpufreq: Kconfig: Fixup incorrect selection by CPU_FREQ_DEFAULT_GOV_SCHED
John Stultz [Thu, 8 Sep 2016 23:43:21 +0000 (16:43 -0700)]
cpufreq: Kconfig: Fixup incorrect selection by CPU_FREQ_DEFAULT_GOV_SCHED

The CPU_FREQ_DEFAULT_GOV_SCHED option is incorrectly selecting
CPU_FREQ_GOV_INTERACTIVE, when it should be selecting
CPU_FREQ_GOV_SCHED.

Signed-off-by: John Stultz <john.stultz@linaro.org>
8 years agoFROMLIST: pstore: drop pmsg bounce buffer
Mark Salyzyn [Wed, 31 Aug 2016 15:09:04 +0000 (08:09 -0700)]
FROMLIST: pstore: drop pmsg bounce buffer

(from https://lkml.org/lkml/2016/9/1/428)
(cherry pick from android-3.10 commit b58133100b38f2bf83cad2d7097417a3a196ed0b)

Removing a bounce buffer copy operation in the pmsg driver path is
always better. We also gain in overall performance by not requesting
a vmalloc on every write as this can cause precious RT tasks, such
as user facing media operation, to stall while memory is being
reclaimed. Added a write_buf_user to the pstore functions, a backup
platform write_buf_user that uses the small buffer that is part of
the instance, and implemented a ramoops write_buf_user that only
supports PSTORE_TYPE_PMSG.

Signed-off-by: Mark Salyzyn <salyzyn@google.com>
Bug: 31057326
Change-Id: I4cdee1cd31467aa3e6c605bce2fbd4de5b0f8caa

8 years agoUPSTREAM: usercopy: remove page-spanning test for now
Kees Cook [Wed, 7 Sep 2016 16:54:34 +0000 (09:54 -0700)]
UPSTREAM: usercopy: remove page-spanning test for now

A custom allocator without __GFP_COMP that copies to userspace has been
found in vmw_execbuf_process[1], so this disables the page-span checker
by placing it behind a CONFIG for future work where such things can be
tracked down later.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=1373326

Reported-by: Vinson Lee <vlee@freedesktop.org>
Fixes: f5509cc18daa ("mm: Hardened usercopy")
Signed-off-by: Kees Cook <keescook@chromium.org>
Change-Id: I4177c0fb943f14a5faf5c70f5e54bf782c316f43
(cherry picked from commit 8e1f74ea02cf4562404c48c6882214821552c13f)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: usercopy: force check_object_size() inline
Kees Cook [Wed, 7 Sep 2016 16:39:32 +0000 (09:39 -0700)]
UPSTREAM: usercopy: force check_object_size() inline

Just for good measure, make sure that check_object_size() is always
inlined too, as already done for copy_*_user() and __copy_*_user().

Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Kees Cook <keescook@chromium.org>
Change-Id: Ibfdf4790d03fe426e68d9a864c55a0d1bbfb7d61
(cherry picked from commit a85d6b8242dc78ef3f4542a0f979aebcbe77fc4e)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoBACKPORT: usercopy: fold builtin_const check into inline function
Kees Cook [Wed, 31 Aug 2016 23:04:21 +0000 (16:04 -0700)]
BACKPORT: usercopy: fold builtin_const check into inline function

Instead of having each caller of check_object_size() need to remember to
check for a const size parameter, move the check into check_object_size()
itself. This actually matches the original implementation in PaX, though
this commit cleans up the now-redundant builtin_const() calls in the
various architectures.

Signed-off-by: Kees Cook <keescook@chromium.org>
Change-Id: I348809399c10ffa051251866063be674d064b9ff
(cherry picked from 81409e9e28058811c9ea865345e1753f8f677e44)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: x86/uaccess: force copy_*_user() to be inlined
Kees Cook [Tue, 6 Sep 2016 18:56:01 +0000 (11:56 -0700)]
UPSTREAM: x86/uaccess: force copy_*_user() to be inlined

As already done with __copy_*_user(), mark copy_*_user() as __always_inline.
Without this, the checks for things like __builtin_const_p() won't work
consistently in either hardened usercopy nor the recent adjustments for
detecting usercopy overflows at compile time.

The change in kernel text size is detectable, but very small:

 text      data     bss     dec      hex     filename
12118735  5768608 14229504 32116847 1ea106f vmlinux.before
12120207  5768608 14229504 32118319 1ea162f vmlinux.after

Signed-off-by: Kees Cook <keescook@chromium.org>
Change-Id: I284c85c2a782145f46655a91d4f83874c90eba61
(cherry picked from commit e6971009a95a74f28c58bbae415c40effad1226c)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: HID: core: prevent out-of-bound readings
Benjamin Tissoires [Tue, 19 Jan 2016 11:34:58 +0000 (12:34 +0100)]
UPSTREAM: HID: core: prevent out-of-bound readings

(cherry picked from commit 50220dead1650609206efe91f0cc116132d59b3f)

Plugging a Logitech DJ receiver with KASAN activated raises a bunch of
out-of-bound readings.

The fields are allocated up to MAX_USAGE, meaning that potentially, we do
not have enough fields to fit the incoming values.
Add checks and silence KASAN.

Signed-off-by: Benjamin Tissoires <benjamin.tissoires@redhat.com>
Signed-off-by: Jiri Kosina <jkosina@suse.cz>
Change-Id: Iaf25e882a6696884439d7091b5fbb0b350d893d3
Bug: 30951261

8 years agoUPSTREAM: block: fix use-after-free in sys_ioprio_get()
Omar Sandoval [Fri, 1 Jul 2016 07:39:35 +0000 (00:39 -0700)]
UPSTREAM: block: fix use-after-free in sys_ioprio_get()

(cherry picked from commit 8ba8682107ee2ca3347354e018865d8e1967c5f4)

get_task_ioprio() accesses the task->io_context without holding the task
lock and thus can race with exit_io_context(), leading to a
use-after-free. The reproducer below hits this within a few seconds on
my 4-core QEMU VM:

#define _GNU_SOURCE
#include <assert.h>
#include <unistd.h>
#include <sys/syscall.h>
#include <sys/wait.h>

int main(int argc, char **argv)
{
pid_t pid, child;
long nproc, i;

/* ioprio_set(IOPRIO_WHO_PROCESS, 0, IOPRIO_PRIO_VALUE(IOPRIO_CLASS_IDLE, 0)); */
syscall(SYS_ioprio_set, 1, 0, 0x6000);

nproc = sysconf(_SC_NPROCESSORS_ONLN);

for (i = 0; i < nproc; i++) {
pid = fork();
assert(pid != -1);
if (pid == 0) {
for (;;) {
pid = fork();
assert(pid != -1);
if (pid == 0) {
_exit(0);
} else {
child = wait(NULL);
assert(child == pid);
}
}
}

pid = fork();
assert(pid != -1);
if (pid == 0) {
for (;;) {
/* ioprio_get(IOPRIO_WHO_PGRP, 0); */
syscall(SYS_ioprio_get, 2, 0);
}
}
}

for (;;) {
/* ioprio_get(IOPRIO_WHO_PGRP, 0); */
syscall(SYS_ioprio_get, 2, 0);
}

return 0;
}

This gets us KASAN dumps like this:

[   35.526914] ==================================================================
[   35.530009] BUG: KASAN: out-of-bounds in get_task_ioprio+0x7b/0x90 at addr ffff880066f34e6c
[   35.530009] Read of size 2 by task ioprio-gpf/363
[   35.530009] =============================================================================
[   35.530009] BUG blkdev_ioc (Not tainted): kasan: bad access detected
[   35.530009] -----------------------------------------------------------------------------

[   35.530009] Disabling lock debugging due to kernel taint
[   35.530009] INFO: Allocated in create_task_io_context+0x2b/0x370 age=0 cpu=0 pid=360
[   35.530009]  ___slab_alloc+0x55d/0x5a0
[   35.530009]  __slab_alloc.isra.20+0x2b/0x40
[   35.530009]  kmem_cache_alloc_node+0x84/0x200
[   35.530009]  create_task_io_context+0x2b/0x370
[   35.530009]  get_task_io_context+0x92/0xb0
[   35.530009]  copy_process.part.8+0x5029/0x5660
[   35.530009]  _do_fork+0x155/0x7e0
[   35.530009]  SyS_clone+0x19/0x20
[   35.530009]  do_syscall_64+0x195/0x3a0
[   35.530009]  return_from_SYSCALL_64+0x0/0x6a
[   35.530009] INFO: Freed in put_io_context+0xe7/0x120 age=0 cpu=0 pid=1060
[   35.530009]  __slab_free+0x27b/0x3d0
[   35.530009]  kmem_cache_free+0x1fb/0x220
[   35.530009]  put_io_context+0xe7/0x120
[   35.530009]  put_io_context_active+0x238/0x380
[   35.530009]  exit_io_context+0x66/0x80
[   35.530009]  do_exit+0x158e/0x2b90
[   35.530009]  do_group_exit+0xe5/0x2b0
[   35.530009]  SyS_exit_group+0x1d/0x20
[   35.530009]  entry_SYSCALL_64_fastpath+0x1a/0xa4
[   35.530009] INFO: Slab 0xffffea00019bcd00 objects=20 used=4 fp=0xffff880066f34ff0 flags=0x1fffe0000004080
[   35.530009] INFO: Object 0xffff880066f34e58 @offset=3672 fp=0x0000000000000001
[   35.530009] ==================================================================

Fix it by grabbing the task lock while we poke at the io_context.

Cc: stable@vger.kernel.org
Reported-by: Dmitry Vyukov <dvyukov@google.com>
Signed-off-by: Omar Sandoval <osandov@fb.com>
Signed-off-by: Jens Axboe <axboe@fb.com>
Change-Id: I3f5858cc9a1b9d4124ae7a6578660dec219d2c57
Bug: 30946378

8 years agoAndroid: Fix build breakages.
Mohan Srinivasan [Thu, 8 Sep 2016 00:39:42 +0000 (17:39 -0700)]
Android: Fix build breakages.

The IO latency histogram change broke allmodconfig and
allnoconfig builds. This fixes those breakages.

Change-Id: I9cdae655b40ed155468f3cef25cdb74bb56c4d3e
Signed-off-by: Mohan Srinivasan <srmohan@google.com>
8 years agoUPSTREAM: tty: Prevent ldisc drivers from re-using stale tty fields
Peter Hurley [Fri, 27 Nov 2015 19:30:21 +0000 (14:30 -0500)]
UPSTREAM: tty: Prevent ldisc drivers from re-using stale tty fields

(cherry picked from commit dd42bf1197144ede075a9d4793123f7689e164bc)

Line discipline drivers may mistakenly misuse ldisc-related fields
when initializing. For example, a failure to initialize tty->receive_room
in the N_GIGASET_M101 line discipline was recently found and fixed [1].
Now, the N_X25 line discipline has been discovered accessing the previous
line discipline's already-freed private data [2].

Harden the ldisc interface against misuse by initializing revelant
tty fields before instancing the new line discipline.

[1]
    commit fd98e9419d8d622a4de91f76b306af6aa627aa9c
    Author: Tilman Schmidt <tilman@imap.cc>
    Date:   Tue Jul 14 00:37:13 2015 +0200

    isdn/gigaset: reset tty->receive_room when attaching ser_gigaset

[2] Report from Sasha Levin <sasha.levin@oracle.com>
    [  634.336761] ==================================================================
    [  634.338226] BUG: KASAN: use-after-free in x25_asy_open_tty+0x13d/0x490 at addr ffff8800a743efd0
    [  634.339558] Read of size 4 by task syzkaller_execu/8981
    [  634.340359] =============================================================================
    [  634.341598] BUG kmalloc-512 (Not tainted): kasan: bad access detected
    ...
    [  634.405018] Call Trace:
    [  634.405277] dump_stack (lib/dump_stack.c:52)
    [  634.405775] print_trailer (mm/slub.c:655)
    [  634.406361] object_err (mm/slub.c:662)
    [  634.406824] kasan_report_error (mm/kasan/report.c:138 mm/kasan/report.c:236)
    [  634.409581] __asan_report_load4_noabort (mm/kasan/report.c:279)
    [  634.411355] x25_asy_open_tty (drivers/net/wan/x25_asy.c:559 (discriminator 1))
    [  634.413997] tty_ldisc_open.isra.2 (drivers/tty/tty_ldisc.c:447)
    [  634.414549] tty_set_ldisc (drivers/tty/tty_ldisc.c:567)
    [  634.415057] tty_ioctl (drivers/tty/tty_io.c:2646 drivers/tty/tty_io.c:2879)
    [  634.423524] do_vfs_ioctl (fs/ioctl.c:43 fs/ioctl.c:607)
    [  634.427491] SyS_ioctl (fs/ioctl.c:622 fs/ioctl.c:613)
    [  634.427945] entry_SYSCALL_64_fastpath (arch/x86/entry/entry_64.S:188)

Cc: Tilman Schmidt <tilman@imap.cc>
Cc: Sasha Levin <sasha.levin@oracle.com>
Signed-off-by: Peter Hurley <peter@hurleysoftware.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Change-Id: Ibed6feadfb9706d478f93feec3b240aecfc64af3
Bug: 30951112

8 years agoUPSTREAM: netfilter: nfnetlink: correctly validate length of batch messages
Phil Turnbull [Tue, 2 Feb 2016 18:36:45 +0000 (13:36 -0500)]
UPSTREAM: netfilter: nfnetlink: correctly validate length of batch messages

(cherry picked from commit c58d6c93680f28ac58984af61d0a7ebf4319c241)

If nlh->nlmsg_len is zero then an infinite loop is triggered because
'skb_pull(skb, msglen);' pulls zero bytes.

The calculation in nlmsg_len() underflows if 'nlh->nlmsg_len <
NLMSG_HDRLEN' which bypasses the length validation and will later
trigger an out-of-bound read.

If the length validation does fail then the malformed batch message is
copied back to userspace. However, we cannot do this because the
nlh->nlmsg_len can be invalid. This leads to an out-of-bounds read in
netlink_ack:

    [   41.455421] ==================================================================
    [   41.456431] BUG: KASAN: slab-out-of-bounds in memcpy+0x1d/0x40 at addr ffff880119e79340
    [   41.456431] Read of size 4294967280 by task a.out/987
    [   41.456431] =============================================================================
    [   41.456431] BUG kmalloc-512 (Not tainted): kasan: bad access detected
    [   41.456431] -----------------------------------------------------------------------------
    ...
    [   41.456431] Bytes b4 ffff880119e79310: 00 00 00 00 d5 03 00 00 b0 fb fe ff 00 00 00 00  ................
    [   41.456431] Object ffff880119e79320: 20 00 00 00 10 00 05 00 00 00 00 00 00 00 00 00   ...............
    [   41.456431] Object ffff880119e79330: 14 00 0a 00 01 03 fc 40 45 56 11 22 33 10 00 05  .......@EV."3...
    [   41.456431] Object ffff880119e79340: f0 ff ff ff 88 99 aa bb 00 14 00 0a 00 06 fe fb  ................
                                            ^^ start of batch nlmsg with
                                               nlmsg_len=4294967280
    ...
    [   41.456431] Memory state around the buggy address:
    [   41.456431]  ffff880119e79400: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    [   41.456431]  ffff880119e79480: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
    [   41.456431] >ffff880119e79500: 00 00 00 00 fc fc fc fc fc fc fc fc fc fc fc fc
    [   41.456431]                                ^
    [   41.456431]  ffff880119e79580: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
    [   41.456431]  ffff880119e79600: fc fc fc fc fc fc fc fc fc fc fb fb fb fb fb fb
    [   41.456431] ==================================================================

Fix this with better validation of nlh->nlmsg_len and by setting
NFNL_BATCH_FAILURE if any batch message fails length validation.

CAP_NET_ADMIN is required to trigger the bugs.

Fixes: 9ea2aa8b7dba ("netfilter: nfnetlink: validate nfnetlink header from batch")
Signed-off-by: Phil Turnbull <phil.turnbull@oracle.com>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>
Change-Id: Id3e15c40cb464bf2791af907c235d8a316b2449c
Bug: 30947055

8 years agocpuset: Make cpusets restore on hotplug
Riley Andrews [Tue, 6 Sep 2016 22:16:25 +0000 (15:16 -0700)]
cpuset: Make cpusets restore on hotplug

This deliberately changes the behavior of the per-cpuset
cpus file to not be effected by hotplug. When a cpu is offlined,
it will be removed from the cpuset/cpus file. When a cpu is onlined,
if the cpuset originally requested that that cpu was part of the cpuset,
that cpu will be restored to the cpuset. The cpus files still
have to be hierachical, but the ranges no longer have to be out of
the currently online cpus, just the physically present cpus.

Change-Id: I22cdf33e7d312117bcefba1aeb0125e1ada289a9
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
8 years agoUPSTREAM: mm/slub: support left redzone
Joonsoo Kim [Tue, 15 Mar 2016 21:55:12 +0000 (14:55 -0700)]
UPSTREAM: mm/slub: support left redzone

SLUB already has a redzone debugging feature.  But it is only positioned
at the end of object (aka right redzone) so it cannot catch left oob.
Although current object's right redzone acts as left redzone of next
object, first object in a slab cannot take advantage of this effect.
This patch explicitly adds a left red zone to each object to detect left
oob more precisely.

Background:

Someone complained to me that left OOB doesn't catch even if KASAN is
enabled which does page allocation debugging.  That page is out of our
control so it would be allocated when left OOB happens and, in this
case, we can't find OOB.  Moreover, SLUB debugging feature can be
enabled without page allocator debugging and, in this case, we will miss
that OOB.

Before trying to implement, I expected that changes would be too
complex, but, it doesn't look that complex to me now.  Almost changes
are applied to debug specific functions so I feel okay.

Signed-off-by: Joonsoo Kim <iamjoonsoo.kim@lge.com>
Cc: Christoph Lameter <cl@linux.com>
Cc: Pekka Enberg <penberg@kernel.org>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Change-Id: Ib893a17ecabd692e6c402e864196bf89cd6781a5
(cherry picked from commit d86bd1bece6fc41d59253002db5441fe960a37f6)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: x86: reorganize SMAP handling in user space accesses
Linus Torvalds [Thu, 17 Dec 2015 17:45:09 +0000 (09:45 -0800)]
UPSTREAM: x86: reorganize SMAP handling in user space accesses

This reorganizes how we do the stac/clac instructions in the user access
code.  Instead of adding the instructions directly to the same inline
asm that does the actual user level access and exception handling, add
them at a higher level.

This is mainly preparation for the next step, where we will expose an
interface to allow users to mark several accesses together as being user
space accesses, but it does already clean up some code:

 - the inlined trivial cases of copy_in_user() now do stac/clac just
   once over the accesses: they used to do one pair around the user
   space read, and another pair around the write-back.

 - the {get,put}_user_ex() macros that are used with the catch/try
   handling don't do any stac/clac at all, because that happens in the
   try/catch surrounding them.

Other than those two cleanups that happened naturally from the
re-organization, this should not make any difference. Yet.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Change-Id: Iaad8756bc8e95876e2e2a7d7bbd333fc478ab441
(cherry picked from commit 11f1a4b9755f5dbc3e822a96502ebe9b044b14d8)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: Make the hardened user-copy code depend on having a hardened allocator
Linus Torvalds [Fri, 19 Aug 2016 19:47:01 +0000 (12:47 -0700)]
UPSTREAM: Make the hardened user-copy code depend on having a hardened allocator

The kernel test robot reported a usercopy failure in the new hardened
sanity checks, due to a page-crossing copy of the FPU state into the
task structure.

This happened because the kernel test robot was testing with SLOB, which
doesn't actually do the required book-keeping for slab allocations, and
as a result the hardening code didn't realize that the task struct
allocation was one single allocation - and the sanity checks fail.

Since SLOB doesn't even claim to support hardening (and you really
shouldn't use it), the straightforward solution is to just make the
usercopy hardening code depend on the allocator supporting it.

Reported-by: kernel test robot <xiaolong.ye@intel.com>
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Change-Id: I37d51f866f873341bf7d5297249899b852e1c6ce
(cherry picked from commit 6040e57658eee6eb1315a26119101ca832d1f854)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoAndroid: MMC/UFS IO Latency Histograms.
Mohan Srinivasan [Fri, 26 Aug 2016 01:31:01 +0000 (18:31 -0700)]
Android: MMC/UFS IO Latency Histograms.

This patch adds a new sysfs node (latency_hist) and reports
IO (svc time) latency histograms. Disabled by default, can be
enabled by echoing 0 into latency_hist, stats can be cleared
by writing 2 into latency_hist. This commit fixes the 32 bit
build breakage in the previous commit. Tested on both 32 bit
and 64 bit arm devices.

Bug: 30677035
Change-Id: I9a615a16616d80f87e75676ac4d078a5c429dcf9
Signed-off-by: Mohan Srinivasan <srmohan@google.com>
8 years agoUPSTREAM: usercopy: fix overlap check for kernel text
Josh Poimboeuf [Mon, 22 Aug 2016 16:53:59 +0000 (11:53 -0500)]
UPSTREAM: usercopy: fix overlap check for kernel text

When running with a local patch which moves the '_stext' symbol to the
very beginning of the kernel text area, I got the following panic with
CONFIG_HARDENED_USERCOPY:

  usercopy: kernel memory exposure attempt detected from ffff88103dfff000 (<linear kernel text>) (4096 bytes)
  ------------[ cut here ]------------
  kernel BUG at mm/usercopy.c:79!
  invalid opcode: 0000 [#1] SMP
  ...
  CPU: 0 PID: 4800 Comm: cp Not tainted 4.8.0-rc3.after+ #1
  Hardware name: Dell Inc. PowerEdge R720/0X3D66, BIOS 2.5.4 01/22/2016
  task: ffff880817444140 task.stack: ffff880816274000
  RIP: 0010:[<ffffffff8121c796>] __check_object_size+0x76/0x413
  RSP: 0018:ffff880816277c40 EFLAGS: 00010246
  RAX: 000000000000006b RBX: ffff88103dfff000 RCX: 0000000000000000
  RDX: 0000000000000000 RSI: ffff88081f80dfa8 RDI: ffff88081f80dfa8
  RBP: ffff880816277c90 R08: 000000000000054c R09: 0000000000000000
  R10: 0000000000000005 R11: 0000000000000006 R12: 0000000000001000
  R13: ffff88103e000000 R14: ffff88103dffffff R15: 0000000000000001
  FS:  00007fb9d1750800(0000) GS:ffff88081f800000(0000) knlGS:0000000000000000
  CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
  CR2: 00000000021d2000 CR3: 000000081a08f000 CR4: 00000000001406f0
  Stack:
   ffff880816277cc8 0000000000010000 000000043de07000 0000000000000000
   0000000000001000 ffff880816277e60 0000000000001000 ffff880816277e28
   000000000000c000 0000000000001000 ffff880816277ce8 ffffffff8136c3a6
  Call Trace:
   [<ffffffff8136c3a6>] copy_page_to_iter_iovec+0xa6/0x1c0
   [<ffffffff8136e766>] copy_page_to_iter+0x16/0x90
   [<ffffffff811970e3>] generic_file_read_iter+0x3e3/0x7c0
   [<ffffffffa06a738d>] ? xfs_file_buffered_aio_write+0xad/0x260 [xfs]
   [<ffffffff816e6262>] ? down_read+0x12/0x40
   [<ffffffffa06a61b1>] xfs_file_buffered_aio_read+0x51/0xc0 [xfs]
   [<ffffffffa06a6692>] xfs_file_read_iter+0x62/0xb0 [xfs]
   [<ffffffff812224cf>] __vfs_read+0xdf/0x130
   [<ffffffff81222c9e>] vfs_read+0x8e/0x140
   [<ffffffff81224195>] SyS_read+0x55/0xc0
   [<ffffffff81003a47>] do_syscall_64+0x67/0x160
   [<ffffffff816e8421>] entry_SYSCALL64_slow_path+0x25/0x25
  RIP: 0033:[<00007fb9d0c33c00>] 0x7fb9d0c33c00
  RSP: 002b:00007ffc9c262f28 EFLAGS: 00000246 ORIG_RAX: 0000000000000000
  RAX: ffffffffffffffda RBX: fffffffffff8ffff RCX: 00007fb9d0c33c00
  RDX: 0000000000010000 RSI: 00000000021c3000 RDI: 0000000000000004
  RBP: 00000000021c3000 R08: 0000000000000000 R09: 00007ffc9c264d6c
  R10: 00007ffc9c262c50 R11: 0000000000000246 R12: 0000000000010000
  R13: 00007ffc9c2630b0 R14: 0000000000000004 R15: 0000000000010000
  Code: 81 48 0f 44 d0 48 c7 c6 90 4d a3 81 48 c7 c0 bb b3 a2 81 48 0f 44 f0 4d 89 e1 48 89 d9 48 c7 c7 68 16 a3 81 31 c0 e8 f4 57 f7 ff <0f> 0b 48 8d 90 00 40 00 00 48 39 d3 0f 83 22 01 00 00 48 39 c3
  RIP  [<ffffffff8121c796>] __check_object_size+0x76/0x413
   RSP <ffff880816277c40>

The checked object's range [ffff88103dfff000ffff88103e000000) is
valid, so there shouldn't have been a BUG.  The hardened usercopy code
got confused because the range's ending address is the same as the
kernel's text starting address at 0xffff88103e000000.  The overlap check
is slightly off.

Fixes: f5509cc18daa ("mm: Hardened usercopy")
Signed-off-by: Josh Poimboeuf <jpoimboe@redhat.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Change-Id: I839dbf4ddbb4d9874026a42abed557eb9b3f8bef
(cherry picked from commit 94cd97af690dd9537818dc9841d0ec68bb1dd877)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: usercopy: avoid potentially undefined behavior in pointer math
Eric Biggers [Fri, 19 Aug 2016 19:15:22 +0000 (12:15 -0700)]
UPSTREAM: usercopy: avoid potentially undefined behavior in pointer math

check_bogus_address() checked for pointer overflow using this expression,
where 'ptr' has type 'const void *':

ptr + n < ptr

Since pointer wraparound is undefined behavior, gcc at -O2 by default
treats it like the following, which would not behave as intended:

(long)n < 0

Fortunately, this doesn't currently happen for kernel code because kernel
code is compiled with -fno-strict-overflow.  But the expression should be
fixed anyway to use well-defined integer arithmetic, since it could be
treated differently by different compilers in the future or could be
reported by tools checking for undefined behavior.

Signed-off-by: Eric Biggers <ebiggers@google.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Change-Id: I73b13be651cf35c03482f2014bf2c3dd291518ab
(cherry picked from commit 7329a655875a2f4bd6984fe8a7e00a6981e802f3)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: mm: SLUB hardened usercopy support
Kees Cook [Thu, 23 Jun 2016 22:24:05 +0000 (15:24 -0700)]
UPSTREAM: mm: SLUB hardened usercopy support

Under CONFIG_HARDENED_USERCOPY, this adds object size checking to the
SLUB allocator to catch any copies that may span objects. Includes a
redzone handling fix discovered by Michael Ellerman.

Based on code from PaX and grsecurity.

Signed-off-by: Kees Cook <keescook@chromium.org>
Tested-by: Michael Ellerman <mpe@ellerman.id.au>
Reviwed-by: Laura Abbott <labbott@redhat.com>
Change-Id: I52dc6fb3a3492b937d52b5cf9c046bf03dc40a3a
(cherry picked from commit ed18adc1cdd00a5c55a20fbdaed4804660772281)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: mm: SLAB hardened usercopy support
Kees Cook [Thu, 23 Jun 2016 22:20:59 +0000 (15:20 -0700)]
UPSTREAM: mm: SLAB hardened usercopy support

Under CONFIG_HARDENED_USERCOPY, this adds object size checking to the
SLAB allocator to catch any copies that may span objects.

Based on code from PaX and grsecurity.

Signed-off-by: Kees Cook <keescook@chromium.org>
Tested-by: Valdis Kletnieks <valdis.kletnieks@vt.edu>
Change-Id: Ib910a71fdc2ab808e1a45b6d33e9bae1681a1f4a
(cherry picked from commit 04385fc5e8fffed84425d909a783c0f0c587d847)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoBACKPORT: arm64/uaccess: Enable hardened usercopy
Kees Cook [Thu, 23 Jun 2016 22:59:42 +0000 (15:59 -0700)]
BACKPORT: arm64/uaccess: Enable hardened usercopy

Enables CONFIG_HARDENED_USERCOPY checks on arm64. As done by KASAN in -next,
renames the low-level functions to __arch_copy_*_user() so a static inline
can do additional work before the copy.

Signed-off-by: Kees Cook <keescook@chromium.org>
Change-Id: I1286cae8e6ffcf12ea54ddd62f1a6d2ce742c8d0
(cherry picked from commit faf5b63e294151d6ac24ca6906d6f221bd3496cd)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoBACKPORT: ARM: uaccess: Enable hardened usercopy
Kees Cook [Thu, 23 Jun 2016 22:06:53 +0000 (15:06 -0700)]
BACKPORT: ARM: uaccess: Enable hardened usercopy

Enables CONFIG_HARDENED_USERCOPY checks on arm.

Based on code from PaX and grsecurity.

Signed-off-by: Kees Cook <keescook@chromium.org>
Change-Id: I03a44ca7a8c56832f15a6a74ac32e9330df3ac3b
(cherry picked from commit dfd45b6103c973bfcea2341d89e36faf947dbc33)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoBACKPORT: x86/uaccess: Enable hardened usercopy
Kees Cook [Thu, 23 Jun 2016 22:04:01 +0000 (15:04 -0700)]
BACKPORT: x86/uaccess: Enable hardened usercopy

Enables CONFIG_HARDENED_USERCOPY checks on x86. This is done both in
copy_*_user() and __copy_*_user() because copy_*_user() actually calls
down to _copy_*_user() and not __copy_*_user().

Based on code from PaX and grsecurity.

Signed-off-by: Kees Cook <keescook@chromium.org>
Tested-by: Valdis Kletnieks <valdis.kletnieks@vt.edu>
Change-Id: I260db1d4572bdd2f779200aca99d03a170658440
(cherry picked from commit 5b710f34e194c6b7710f69fdb5d798fdf35b98c1)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoBACKPORT: mm: Hardened usercopy
Kees Cook [Tue, 7 Jun 2016 18:05:33 +0000 (11:05 -0700)]
BACKPORT: mm: Hardened usercopy

This is the start of porting PAX_USERCOPY into the mainline kernel. This
is the first set of features, controlled by CONFIG_HARDENED_USERCOPY. The
work is based on code by PaX Team and Brad Spengler, and an earlier port
from Casey Schaufler. Additional non-slab page tests are from Rik van Riel.

This patch contains the logic for validating several conditions when
performing copy_to_user() and copy_from_user() on the kernel object
being copied to/from:
- address range doesn't wrap around
- address range isn't NULL or zero-allocated (with a non-zero copy size)
- if on the slab allocator:
  - object size must be less than or equal to copy size (when check is
    implemented in the allocator, which appear in subsequent patches)
- otherwise, object must not span page allocations (excepting Reserved
  and CMA ranges)
- if on the stack
  - object must not extend before/after the current process stack
  - object must be contained by a valid stack frame (when there is
    arch/build support for identifying stack frames)
- object must not overlap with kernel text

Signed-off-by: Kees Cook <keescook@chromium.org>
Tested-by: Valdis Kletnieks <valdis.kletnieks@vt.edu>
Tested-by: Michael Ellerman <mpe@ellerman.id.au>
Change-Id: Iff3b5f1ddb04acd99ccf9a9046c7797363962b2a
(cherry picked from commit f5509cc18daa7f82bcc553be70df2117c8eedc16)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoBACKPORT: mm: Implement stack frame object validation
Kees Cook [Tue, 12 Jul 2016 23:19:48 +0000 (16:19 -0700)]
BACKPORT: mm: Implement stack frame object validation

This creates per-architecture function arch_within_stack_frames() that
should validate if a given object is contained by a kernel stack frame.
Initial implementation is on x86.

This is based on code from PaX.

Signed-off-by: Kees Cook <keescook@chromium.org>
Change-Id: I1f3b299bb8991d65dcdac6af85d633d4b7776df1
(cherry picked from commit 0f60a8efe4005ab5e65ce000724b04d4ca04a199)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: mm: Add is_migrate_cma_page
Laura Abbott [Tue, 19 Jul 2016 22:00:04 +0000 (15:00 -0700)]
UPSTREAM: mm: Add is_migrate_cma_page

Code such as hardened user copy[1] needs a way to tell if a
page is CMA or not. Add is_migrate_cma_page in a similar way
to is_migrate_isolate_page.

[1]http://article.gmane.org/gmane.linux.kernel.mm/155238

Signed-off-by: Laura Abbott <labbott@redhat.com>
Signed-off-by: Kees Cook <keescook@chromium.org>
Change-Id: I1f9aa13d8d063038fa70b93282a836648fbb4f6d
(cherry picked from commit 7c15d9bb8231f998ae7dc0b72415f5215459f7fb)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: unsafe_[get|put]_user: change interface to use a error target label
Linus Torvalds [Mon, 8 Aug 2016 20:02:01 +0000 (13:02 -0700)]
UPSTREAM: unsafe_[get|put]_user: change interface to use a error target label

When I initially added the unsafe_[get|put]_user() helpers in commit
5b24a7a2aa20 ("Add 'unsafe' user access functions for batched
accesses"), I made the mistake of modeling the interface on our
traditional __[get|put]_user() functions, which return zero on success,
or -EFAULT on failure.

That interface is fairly easy to use, but it's actually fairly nasty for
good code generation, since it essentially forces the caller to check
the error value for each access.

In particular, since the error handling is already internally
implemented with an exception handler, and we already use "asm goto" for
various other things, we could fairly easily make the error cases just
jump directly to an error label instead, and avoid the need for explicit
checking after each operation.

So switch the interface to pass in an error label, rather than checking
the error value in the caller.  Best do it now before we start growing
more users (the signal handling code in particular would be a good place
to use the new interface).

So rather than

if (unsafe_get_user(x, ptr))
... handle error ..

the interface is now

unsafe_get_user(x, ptr, label);

where an error during the user mode fetch will now just cause a jump to
'label' in the caller.

Right now the actual _implementation_ of this all still ends up being a
"if (err) goto label", and does not take advantage of any exception
label tricks, but for "unsafe_put_user()" in particular it should be
fairly straightforward to convert to using the exception table model.

Note that "unsafe_get_user()" is much harder to convert to a clever
exception table model, because current versions of gcc do not allow the
use of "asm goto" (for the exception) with output values (for the actual
value to be fetched).  But that is hopefully not a limitation in the
long term.

[ Also note that it might be a good idea to switch unsafe_get_user() to
  actually _return_ the value it fetches from user space, but this
  commit only changes the error handling semantics ]

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Change-Id: Ib905a84a04d46984320f6fd1056da4d72f3d6b53
(cherry picked from commit 1bd4403d86a1c06cb6cc9ac87664a0c9d3413d51)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoBACKPORT: arm64: mm: fix location of _etext
Ard Biesheuvel [Thu, 23 Jun 2016 13:53:17 +0000 (15:53 +0200)]
BACKPORT: arm64: mm: fix location of _etext

As Kees Cook notes in the ARM counterpart of this patch [0]:

  The _etext position is defined to be the end of the kernel text code,
  and should not include any part of the data segments. This interferes
  with things that might check memory ranges and expect executable code
  up to _etext.

In particular, Kees is referring to the HARDENED_USERCOPY patch set [1],
which rejects attempts to call copy_to_user() on kernel ranges containing
executable code, but does allow access to the .rodata segment. Regardless
of whether one may or may not agree with the distinction, it makes sense
for _etext to have the same meaning across architectures.

So let's put _etext where it belongs, between .text and .rodata, and fix
up existing references to use __init_begin instead, which unlike _end_rodata
includes the exception and notes sections as well.

The _etext references in kaslr.c are left untouched, since its references
to [_stext, _etext) are meant to capture potential jump instruction targets,
and so disregarding .rodata is actually an improvement here.

[0] http://article.gmane.org/gmane.linux.kernel/2245084
[1] http://thread.gmane.org/gmane.linux.kernel.hardened.devel/2502

Reported-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Mark Rutland <mark.rutland@arm.com>
Signed-off-by: Ard Biesheuvel <ard.biesheuvel@linaro.org>
Reviewed-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Change-Id: I8f6582525217b9ca324f6a382ea52d30ce1d0dbd
(cherry picked from commit 9fdc14c55cd6579d619ccd9d40982e0805e62b6d)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoBACKPORT: ARM: 8583/1: mm: fix location of _etext
Kees Cook [Thu, 23 Jun 2016 20:28:47 +0000 (21:28 +0100)]
BACKPORT: ARM: 8583/1: mm: fix location of _etext

The _etext position is defined to be the end of the kernel text code,
and should not include any part of the data segments. This interferes
with things that might check memory ranges and expect executable code
up to _etext. Just to be conservative, leave the kernel resource as
it was, using __init_begin instead of _etext as the end mark.

Signed-off-by: Kees Cook <keescook@chromium.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
Change-Id: Ida514d1359dbe6f782f562ce29b4ba09ae72bfc0
(cherry picked from commit 14c4a533e0996f95a0a64dfd0b6252d788cebc74)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
8 years agoUPSTREAM: Use the new batched user accesses in generic user string handling
Linus Torvalds [Thu, 17 Dec 2015 18:05:19 +0000 (10:05 -0800)]
UPSTREAM: Use the new batched user accesses in generic user string handling

This converts the generic user string functions to use the batched user
access functions.

It makes a big difference on Skylake, which is the first x86
microarchitecture to implement SMAP.  The STAC/CLAC instructions are not
very fast, and doing them for each access inside the loop that copies
strings from user space (which is what the pathname handling does for
every pathname the kernel uses, for example) is very inefficient.

Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
Change-Id: Ic39a686b4bb1ad9cd16ad0887bb669beed6fe8aa
(cherry picked from commit 9fd4470ff4974c41b1db43c3b355b9085af9c12a)
Signed-off-by: Sami Tolvanen <samitolvanen@google.com>