platform/kernel/linux-exynos.git
8 years agofs/cifs: correctly to anonymous authentication for the LANMAN authentication
Stefan Metzmacher [Tue, 3 May 2016 08:52:30 +0000 (10:52 +0200)]
fs/cifs: correctly to anonymous authentication for the LANMAN authentication

[ Upstream commit fa8f3a354bb775ec586e4475bcb07f7dece97e0c ]

Only server which map unknown users to guest will allow
access using a non-null LMChallengeResponse.

For Samba it's the "map to guest = bad user" option.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=11913

Signed-off-by: Stefan Metzmacher <metze@samba.org>
CC: Stable <stable@vger.kernel.org>
Signed-off-by: Steve French <smfrench@gmail.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agofs/cifs: correctly to anonymous authentication via NTLMSSP
Stefan Metzmacher [Tue, 3 May 2016 08:52:30 +0000 (10:52 +0200)]
fs/cifs: correctly to anonymous authentication via NTLMSSP

[ Upstream commit cfda35d98298131bf38fbad3ce4cd5ecb3cf18db ]

See [MS-NLMP] 3.2.5.1.2 Server Receives an AUTHENTICATE_MESSAGE from the Client:

   ...
   Set NullSession to FALSE
   If (AUTHENTICATE_MESSAGE.UserNameLen == 0 AND
      AUTHENTICATE_MESSAGE.NtChallengeResponse.Length == 0 AND
      (AUTHENTICATE_MESSAGE.LmChallengeResponse == Z(1)
       OR
       AUTHENTICATE_MESSAGE.LmChallengeResponse.Length == 0))
       -- Special case: client requested anonymous authentication
       Set NullSession to TRUE
   ...

Only server which map unknown users to guest will allow
access using a non-null NTChallengeResponse.

For Samba it's the "map to guest = bad user" option.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=11913

CC: Stable <stable@vger.kernel.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Steve French <smfrench@gmail.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agodrm/fb_helper: Fix references to dev->mode_config.num_connector
Lyude [Thu, 12 May 2016 14:56:59 +0000 (10:56 -0400)]
drm/fb_helper: Fix references to dev->mode_config.num_connector

[ Upstream commit 255f0e7c418ad95a4baeda017ae6182ba9b3c423 ]

During boot, MST hotplugs are generally expected (even if no physical
hotplugging occurs) and result in DRM's connector topology changing.
This means that using num_connector from the current mode configuration
can lead to the number of connectors changing under us. This can lead to
some nasty scenarios in fbcon:

- We allocate an array to the size of dev->mode_config.num_connectors.
- MST hotplug occurs, dev->mode_config.num_connectors gets incremented.
- We try to loop through each element in the array using the new value
  of dev->mode_config.num_connectors, and end up going out of bounds
  since dev->mode_config.num_connectors is now larger then the array we
  allocated.

fb_helper->connector_count however, will always remain consistent while
we do a modeset in fb_helper.

Note: This is just polish for 4.7, Dave Airlie's drm_connector
refcounting fixed these bugs for real. But it's good enough duct-tape
for stable kernel backporting, since backporting the refcounting
changes is way too invasive.

Cc: stable@vger.kernel.org
Signed-off-by: Lyude <cpaul@redhat.com>
[danvet: Clarify why we need this. Also remove the now unused "dev"
local variable to appease gcc.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1463065021-18280-3-git-send-email-cpaul@redhat.com
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agodrm/i915/fbdev: Fix num_connector references in intel_fb_initial_config()
Lyude [Thu, 12 May 2016 14:56:58 +0000 (10:56 -0400)]
drm/i915/fbdev: Fix num_connector references in intel_fb_initial_config()

[ Upstream commit 14a3842a1d5945067d1dd0788f314e14d5b18e5b ]

During boot time, MST devices usually send a ton of hotplug events
irregardless of whether or not any physical hotplugs actually occurred.
Hotplugs mean connectors being created/destroyed, and the number of DRM
connectors changing under us. This isn't a problem if we use
fb_helper->connector_count since we only set it once in the code,
however if we use num_connector from struct drm_mode_config we risk it's
value changing under us. On top of that, there's even a chance that
dev->mode_config.num_connector != fb_helper->connector_count. If the
number of connectors happens to increase under us, we'll end up using
the wrong array size for memcpy and start writing beyond the actual
length of the array, occasionally resulting in kernel panics.

Note: This is just polish for 4.7, Dave Airlie's drm_connector
refcounting fixed these bugs for real. But it's good enough duct-tape
for stable kernel backporting, since backporting the refcounting
changes is way too invasive.

Cc: stable@vger.kernel.org
Signed-off-by: Lyude <cpaul@redhat.com>
[danvet: Clarify why we need this.]
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/1463065021-18280-2-git-send-email-cpaul@redhat.com
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: MSA: Fix a link error on `_init_msa_upper' with older GCC
Maciej W. Rozycki [Tue, 17 May 2016 05:12:27 +0000 (06:12 +0100)]
MIPS: MSA: Fix a link error on `_init_msa_upper' with older GCC

[ Upstream commit e49d38488515057dba8f0c2ba4cfde5be4a7281f ]

Fix a build regression from commit c9017757c532 ("MIPS: init upper 64b
of vector registers when MSA is first used"):

arch/mips/built-in.o: In function `enable_restore_fp_context':
traps.c:(.text+0xbb90): undefined reference to `_init_msa_upper'
traps.c:(.text+0xbb90): relocation truncated to fit: R_MIPS_26 against `_init_msa_upper'
traps.c:(.text+0xbef0): undefined reference to `_init_msa_upper'
traps.c:(.text+0xbef0): relocation truncated to fit: R_MIPS_26 against `_init_msa_upper'

to !CONFIG_CPU_HAS_MSA configurations with older GCC versions, which are
unable to figure out that calls to `_init_msa_upper' are indeed dead.
Of the many ways to tackle this failure choose the approach we have
already taken in `thread_msa_context_live'.

[ralf@linux-mips.org: Drop patch segment to junk file.]

Signed-off-by: Maciej W. Rozycki <macro@imgtec.com>
Cc: stable@vger.kernel.org # v3.16+
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/13271/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoPCI: Disable all BAR sizing for devices with non-compliant BARs
Prarit Bhargava [Wed, 11 May 2016 16:27:16 +0000 (12:27 -0400)]
PCI: Disable all BAR sizing for devices with non-compliant BARs

[ Upstream commit ad67b437f187ea818b2860524d10f878fadfdd99 ]

b84106b4e229 ("PCI: Disable IO/MEM decoding for devices with non-compliant
BARs") disabled BAR sizing for BARs 0-5 of devices that don't comply with
the PCI spec.  But it didn't do anything for expansion ROM BARs, so we
still try to size them, resulting in warnings like this on Broadwell-EP:

  pci 0000:ff:12.0: BAR 6: failed to assign [mem size 0x00000001 pref]

Move the non-compliant BAR check from __pci_read_base() up to
pci_read_bases() so it applies to the expansion ROM BAR as well as
to BARs 0-5.

Note that direct callers of __pci_read_base(), like sriov_init(), will now
bypass this check.  We haven't had reports of devices with broken SR-IOV
BARs yet.

[bhelgaas: changelog]
Fixes: b84106b4e229 ("PCI: Disable IO/MEM decoding for devices with non-compliant BARs")
Signed-off-by: Prarit Bhargava <prarit@redhat.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: stable@vger.kernel.org
CC: Thomas Gleixner <tglx@linutronix.de>
CC: Ingo Molnar <mingo@redhat.com>
CC: "H. Peter Anvin" <hpa@zytor.com>
CC: Andi Kleen <ak@linux.intel.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agommc: mmc: Fix partition switch timeout for some eMMCs
Adrian Hunter [Thu, 5 May 2016 05:12:28 +0000 (08:12 +0300)]
mmc: mmc: Fix partition switch timeout for some eMMCs

[ Upstream commit 1c447116d017a98c90f8f71c8c5a611e0aa42178 ]

Some eMMCs set the partition switch timeout too low.

Now typically eMMCs are considered a critical component (e.g. because
they store the root file system) and consequently are expected to be
reliable.  Thus we can neglect the use case where eMMCs can't switch
reliably and we might want a lower timeout to facilitate speedy
recovery.

Although we could employ a quirk for the cards that are affected (if
we could identify them all), as described above, there is little
benefit to having a low timeout, so instead simply set a minimum
timeout.

The minimum is set to 300ms somewhat arbitrarily - the examples that
have been seen had a timeout of 10ms but were sometimes taking 60-70ms.

Cc: stable@vger.kernel.org
Signed-off-by: Adrian Hunter <adrian.hunter@intel.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoring-buffer: Prevent overflow of size in ring_buffer_resize()
Steven Rostedt (Red Hat) [Fri, 13 May 2016 13:34:12 +0000 (09:34 -0400)]
ring-buffer: Prevent overflow of size in ring_buffer_resize()

[ Upstream commit 59643d1535eb220668692a5359de22545af579f6 ]

If the size passed to ring_buffer_resize() is greater than MAX_LONG - BUF_PAGE_SIZE
then the DIV_ROUND_UP() will return zero.

Here's the details:

  # echo 18014398509481980 > /sys/kernel/debug/tracing/buffer_size_kb

tracing_entries_write() processes this and converts kb to bytes.

 18014398509481980 << 10 = 18446744073709547520

and this is passed to ring_buffer_resize() as unsigned long size.

 size = DIV_ROUND_UP(size, BUF_PAGE_SIZE);

Where DIV_ROUND_UP(a, b) is (a + b - 1)/b

BUF_PAGE_SIZE is 4080 and here

 18446744073709547520 + 4080 - 1 = 18446744073709551599

where 18446744073709551599 is still smaller than 2^64

 2^64 - 18446744073709551599 = 17

But now 18446744073709551599 / 4080 = 4521260802379792

and size = size * 4080 = 18446744073709551360

This is checked to make sure its still greater than 2 * 4080,
which it is.

Then we convert to the number of buffer pages needed.

 nr_page = DIV_ROUND_UP(size, BUF_PAGE_SIZE)

but this time size is 18446744073709551360 and

 2^64 - (18446744073709551360 + 4080 - 1) = -3823

Thus it overflows and the resulting number is less than 4080, which makes

  3823 / 4080 = 0

an nr_pages is set to this. As we already checked against the minimum that
nr_pages may be, this causes the logic to fail as well, and we crash the
kernel.

There's no reason to have the two DIV_ROUND_UP() (that's just result of
historical code changes), clean up the code and fix this bug.

Cc: stable@vger.kernel.org # 3.5+
Fixes: 83f40318dab00 ("ring-buffer: Make removal of ring buffer pages atomic")
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoring-buffer: Use long for nr_pages to avoid overflow failures
Steven Rostedt (Red Hat) [Thu, 12 May 2016 15:01:24 +0000 (11:01 -0400)]
ring-buffer: Use long for nr_pages to avoid overflow failures

[ Upstream commit 9b94a8fba501f38368aef6ac1b30e7335252a220 ]

The size variable to change the ring buffer in ftrace is a long. The
nr_pages used to update the ring buffer based on the size is int. On 64 bit
machines this can cause an overflow problem.

For example, the following will cause the ring buffer to crash:

 # cd /sys/kernel/debug/tracing
 # echo 10 > buffer_size_kb
 # echo 8556384240 > buffer_size_kb

Then you get the warning of:

 WARNING: CPU: 1 PID: 318 at kernel/trace/ring_buffer.c:1527 rb_update_pages+0x22f/0x260

Which is:

  RB_WARN_ON(cpu_buffer, nr_removed);

Note each ring buffer page holds 4080 bytes.

This is because:

 1) 10 causes the ring buffer to have 3 pages.
    (10kb requires 3 * 4080 pages to hold)

 2) (2^31 / 2^10  + 1) * 4080 = 8556384240
    The value written into buffer_size_kb is shifted by 10 and then passed
    to ring_buffer_resize(). 8556384240 * 2^10 = 8761737461760

 3) The size passed to ring_buffer_resize() is then divided by BUF_PAGE_SIZE
    which is 4080. 8761737461760 / 4080 = 2147484672

 4) nr_pages is subtracted from the current nr_pages (3) and we get:
    2147484669. This value is saved in a signed integer nr_pages_to_update

 5) 2147484669 is greater than 2^31 but smaller than 2^32, a signed int
    turns into the value of -2147482627

 6) As the value is a negative number, in update_pages_handler() it is
    negated and passed to rb_remove_pages() and 2147482627 pages will
    be removed, which is much larger than 3 and it causes the warning
    because not all the pages asked to be removed were removed.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=118001
Cc: stable@vger.kernel.org # 2.6.28+
Fixes: 7a8e76a3829f1 ("tracing: unified trace buffer")
Reported-by: Hao Qin <QEver.cn@gmail.com>
Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoring-buffer: Move recursive check to per_cpu descriptor
Steven Rostedt (Red Hat) [Wed, 27 May 2015 14:27:47 +0000 (10:27 -0400)]
ring-buffer: Move recursive check to per_cpu descriptor

[ Upstream commit 58a09ec6e3ec88c9c7e061479f1ef7fe93324a87 ]

Instead of using a global per_cpu variable to perform the recursive
checks into the ring buffer, use the already existing per_cpu descriptor
that is part of the ring buffer itself.

Not only does this simplify the code, it also allows for one ring buffer
to be used within the guts of the use of another ring buffer. For example
trace_printk() can now be used within the ring buffer to record changes
done by an instance into the main ring buffer. The recursion checks
will prevent the trace_printk() itself from causing recursive issues
with the main ring buffer (it is just ignored), but the recursive
checks wont prevent the trace_printk() from recording other ring buffers.

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoring-buffer: Add unlikelys to make fast path the default
Steven Rostedt (Red Hat) [Thu, 21 May 2015 21:39:29 +0000 (17:39 -0400)]
ring-buffer: Add unlikelys to make fast path the default

[ Upstream commit 3205f8063b6cc54b20d5080fb79dfcbd9c39e93d ]

I was running the trace_event benchmark and noticed that the times
to record a trace_event was all over the place. I looked at the assembly
of the ring_buffer_lock_reserver() and saw this:

 <ring_buffer_lock_reserve>:
       31 c0                   xor    %eax,%eax
       48 83 3d 76 47 bd 00    cmpq   $0x1,0xbd4776(%rip)        # ffffffff81d10d60 <ring_buffer_flags>
       01
       55                      push   %rbp
       48 89 e5                mov    %rsp,%rbp
       75 1d                   jne    ffffffff8113c60d <ring_buffer_lock_reserve+0x2d>
       65 ff 05 69 e3 ec 7e    incl   %gs:0x7eece369(%rip)        # a960 <__preempt_count>
       8b 47 08                mov    0x8(%rdi),%eax
       85 c0                   test   %eax,%eax
 +---- 74 12                   je     ffffffff8113c610 <ring_buffer_lock_reserve+0x30>
 |     65 ff 0d 5b e3 ec 7e    decl   %gs:0x7eece35b(%rip)        # a960 <__preempt_count>
 |     0f 84 85 00 00 00       je     ffffffff8113c690 <ring_buffer_lock_reserve+0xb0>
 |     31 c0                   xor    %eax,%eax
 |     5d                      pop    %rbp
 |     c3                      retq
 |     90                      nop
 +---> 65 44 8b 05 48 e3 ec    mov    %gs:0x7eece348(%rip),%r8d        # a960 <__preempt_count>
       7e
       41 81 e0 ff ff ff 7f    and    $0x7fffffff,%r8d
       b0 08                   mov    $0x8,%al
       65 8b 0d 58 36 ed 7e    mov    %gs:0x7eed3658(%rip),%ecx        # fc80 <current_context>
       41 f7 c0 00 ff 1f 00    test   $0x1fff00,%r8d
       74 1e                   je     ffffffff8113c64f <ring_buffer_lock_reserve+0x6f>
       41 f7 c0 00 00 10 00    test   $0x100000,%r8d
       b0 01                   mov    $0x1,%al
       75 13                   jne    ffffffff8113c64f <ring_buffer_lock_reserve+0x6f>
       41 81 e0 00 00 0f 00    and    $0xf0000,%r8d
       49 83 f8 01             cmp    $0x1,%r8
       19 c0                   sbb    %eax,%eax
       83 e0 02                and    $0x2,%eax
       83 c0 02                add    $0x2,%eax
       85 c8                   test   %ecx,%eax
       75 ab                   jne    ffffffff8113c5fe <ring_buffer_lock_reserve+0x1e>
       09 c8                   or     %ecx,%eax
       65 89 05 24 36 ed 7e    mov    %eax,%gs:0x7eed3624(%rip)        # fc80 <current_context>

The arrow is the fast path.

After adding the unlikely's, the fast path looks a bit better:

 <ring_buffer_lock_reserve>:
       31 c0                   xor    %eax,%eax
       48 83 3d 76 47 bd 00    cmpq   $0x1,0xbd4776(%rip)        # ffffffff81d10d60 <ring_buffer_flags>
       01
       55                      push   %rbp
       48 89 e5                mov    %rsp,%rbp
       75 7b                   jne    ffffffff8113c66b <ring_buffer_lock_reserve+0x8b>
       65 ff 05 69 e3 ec 7e    incl   %gs:0x7eece369(%rip)        # a960 <__preempt_count>
       8b 47 08                mov    0x8(%rdi),%eax
       85 c0                   test   %eax,%eax
       0f 85 9f 00 00 00       jne    ffffffff8113c6a1 <ring_buffer_lock_reserve+0xc1>
       65 8b 0d 57 e3 ec 7e    mov    %gs:0x7eece357(%rip),%ecx        # a960 <__preempt_count>
       81 e1 ff ff ff 7f       and    $0x7fffffff,%ecx
       b0 08                   mov    $0x8,%al
       65 8b 15 68 36 ed 7e    mov    %gs:0x7eed3668(%rip),%edx        # fc80 <current_context>
       f7 c1 00 ff 1f 00       test   $0x1fff00,%ecx
       75 50                   jne    ffffffff8113c670 <ring_buffer_lock_reserve+0x90>
       85 d0                   test   %edx,%eax
       75 7d                   jne    ffffffff8113c6a1 <ring_buffer_lock_reserve+0xc1>
       09 d0                   or     %edx,%eax
       65 89 05 53 36 ed 7e    mov    %eax,%gs:0x7eed3653(%rip)        # fc80 <current_context>
       65 8b 05 fc da ec 7e    mov    %gs:0x7eecdafc(%rip),%eax        # a130 <cpu_number>
       89 c2                   mov    %eax,%edx

Signed-off-by: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: Disable preemption during prctl(PR_SET_FP_MODE, ...)
Paul Burton [Thu, 21 Apr 2016 11:43:57 +0000 (12:43 +0100)]
MIPS: Disable preemption during prctl(PR_SET_FP_MODE, ...)

[ Upstream commit bd239f1e1429e7781096bf3884bdb1b2b1bb4f28 ]

Whilst a PR_SET_FP_MODE prctl is performed there are decisions made
based upon whether the task is executing on the current CPU. This may
change if we're preempted, so disable preemption to avoid such changes
for the lifetime of the mode switch.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Fixes: 9791554b45a2 ("MIPS,prctl: add PR_[GS]ET_FP_MODE prctl options for MIPS")
Reviewed-by: Maciej W. Rozycki <macro@imgtec.com>
Tested-by: Aurelien Jarno <aurelien@aurel32.net>
Cc: Adam Buchbinder <adam.buchbinder@gmail.com>
Cc: James Hogan <james.hogan@imgtec.com>
Cc: stable <stable@vger.kernel.org> # v4.0+
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/13144/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: ptrace: Prevent writes to read-only FCSR bits
Maciej W. Rozycki [Thu, 12 May 2016 09:19:08 +0000 (10:19 +0100)]
MIPS: ptrace: Prevent writes to read-only FCSR bits

[ Upstream commit abf378be49f38c4d3e23581d3df3fa9f1b1b11d2 ]

Correct the cases missed with commit 9b26616c8d9d ("MIPS: Respect the
ISA level in FCSR handling") and prevent writes to read-only FCSR bits
there.

This in particular applies to FP context initialisation where any IEEE
754-2008 bits preset by `mips_set_personality_nan' are cleared before
the relevant ptrace(2) call takes effect and the PTRACE_POKEUSR request
addressing FPC_CSR where no masking of read-only FCSR bits is done.

Remove the FCSR clearing from FP context initialisation then and unify
PTRACE_POKEUSR/FPC_CSR and PTRACE_SETFPREGS handling, by factoring out
code from `ptrace_setfpregs' and calling it from both places.

This mostly matters to soft float configurations where the emulator can
be switched this way to a mode which should not be accessible and cannot
be set with the CTC1 instruction.  With hard float configurations any
effect is transient anyway as read-only bits will retain their values at
the time the FP context is restored.

Signed-off-by: Maciej W. Rozycki <macro@imgtec.com>
Cc: stable@vger.kernel.org # v4.0+
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/13239/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: ptrace: Fix FP context restoration FCSR regression
Maciej W. Rozycki [Thu, 12 May 2016 09:18:27 +0000 (10:18 +0100)]
MIPS: ptrace: Fix FP context restoration FCSR regression

[ Upstream commit 4249548454f7ba4581aeee26bd83f42b48a14d15 ]

Fix a floating-point context restoration regression introduced with
commit 9b26616c8d9d ("MIPS: Respect the ISA level in FCSR handling")
that causes a Floating Point exception and consequently a kernel oops
with hard float configurations when one or more FCSR Enable and their
corresponding Cause bits are set both at a time via a ptrace(2) call.

To do so reinstate Cause bit masking originally introduced with commit
b1442d39fac2 ("MIPS: Prevent user from setting FCSR cause bits") to
address this exact problem and then inadvertently removed from the
PTRACE_SETFPREGS request with the commit referred above.

Signed-off-by: Maciej W. Rozycki <macro@imgtec.com>
Cc: stable@vger.kernel.org # v4.0+
Cc: linux-mips@linux-mips.org
Patchwork: https://patchwork.linux-mips.org/patch/13238/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: math-emu: Fix jalr emulation when rd == $0
Paul Burton [Thu, 21 Apr 2016 13:04:55 +0000 (14:04 +0100)]
MIPS: math-emu: Fix jalr emulation when rd == $0

[ Upstream commit ab4a92e66741b35ca12f8497896bafbe579c28a1 ]

When emulating a jalr instruction with rd == $0, the code in
isBranchInstr was incorrectly writing to GPR $0 which should actually
always remain zeroed. This would lead to any further instructions
emulated which use $0 operating on a bogus value until the task is next
context switched, at which point the value of $0 in the task context
would be restored to the correct zero by a store in SAVE_SOME. Fix this
by not writing to rd if it is $0.

Fixes: 102cedc32a6e ("MIPS: microMIPS: Floating point support.")
Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: Maciej W. Rozycki <macro@imgtec.com>
Cc: James Hogan <james.hogan@imgtec.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Cc: stable <stable@vger.kernel.org> # v3.10
Patchwork: https://patchwork.linux-mips.org/patch/13160/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: Fix uapi include in exported asm/siginfo.h
James Hogan [Mon, 8 Feb 2016 18:43:51 +0000 (18:43 +0000)]
MIPS: Fix uapi include in exported asm/siginfo.h

[ Upstream commit 987e5b834467c9251ca584febda65ef8f66351a9 ]

Since commit 8cb48fe169dd ("MIPS: Provide correct siginfo_t.si_stime"),
MIPS' uapi/asm/siginfo.h has included uapi/asm-generic/siginfo.h
directly before defining MIPS' struct siginfo, in order to get the
necessary definitions needed for the siginfo struct without the generic
copy_siginfo() hitting compiler errors due to struct siginfo not yet
being defined.

Now that the generic copy_siginfo() is moved out to linux/signal.h we
can safely include asm-generic/siginfo.h before defining the MIPS
specific struct siginfo, which avoids the uapi/ include as well as
breakage due to generic copy_siginfo() being defined before struct
siginfo.

Reported-by: Christopher Ferris <cferris@google.com>
Fixes: 8cb48fe169dd ("MIPS: Provide correct siginfo_t.si_stime")
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Petr Malat <oss@malat.biz>
Cc: linux-mips@linux-mips.org
Cc: <stable@vger.kernel.org> # 4.0-
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoSIGNAL: Move generic copy_siginfo() to signal.h
James Hogan [Mon, 8 Feb 2016 18:43:50 +0000 (18:43 +0000)]
SIGNAL: Move generic copy_siginfo() to signal.h

[ Upstream commit ca9eb49aa9562eaadf3cea071ec7018ad6800425 ]

The generic copy_siginfo() is currently defined in
asm-generic/siginfo.h, after including uapi/asm-generic/siginfo.h which
defines the generic struct siginfo. However this makes it awkward for an
architecture to use it if it has to define its own struct siginfo (e.g.
MIPS and potentially IA64), since it means that asm-generic/siginfo.h
can only be included after defining the arch-specific siginfo, which may
be problematic if the arch-specific definition needs definitions from
uapi/asm-generic/siginfo.h.

It is possible to work around this by first including
uapi/asm-generic/siginfo.h to get the constants before defining the
arch-specific siginfo, and include asm-generic/siginfo.h after. However
uapi headers can't be included by other uapi headers, so that first
include has to be in an ifdef __kernel__, with the non __kernel__ case
including the non-UAPI header instead.

Instead of that mess, move the generic copy_siginfo() definition into
linux/signal.h, which allows an arch-specific uapi/asm/siginfo.h to
include asm-generic/siginfo.h and define the arch-specific siginfo, and
for the generic copy_siginfo() to see that arch-specific definition.

Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Arnd Bergmann <arnd@arndb.de>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: Petr Malat <oss@malat.biz>
Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Christopher Ferris <cferris@google.com>
Cc: linux-arch@vger.kernel.org
Cc: linux-mips@linux-mips.org
Cc: linux-ia64@vger.kernel.org
Cc: linux-kernel@vger.kernel.org
Cc: <stable@vger.kernel.org> # 4.0-
Patchwork: https://patchwork.linux-mips.org/patch/12478/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: Sync icache & dcache in set_pte_at
Paul Burton [Tue, 1 Mar 2016 02:37:59 +0000 (02:37 +0000)]
MIPS: Sync icache & dcache in set_pte_at

[ Upstream commit 37d22a0d798b5c938b277d32cfd86dc231381342 ]

It's possible for pages to become visible prior to update_mmu_cache
running if a thread within the same address space preempts the current
thread or runs simultaneously on another CPU. That is, the following
scenario is possible:

    CPU0                            CPU1

    write to page
    flush_dcache_page
    flush_icache_page
    set_pte_at
                                    map page
    update_mmu_cache

If CPU1 maps the page in between CPU0's set_pte_at, which marks it valid
& visible, and update_mmu_cache where the dcache flush occurs then CPU1s
icache will fill from stale data (unless it fills from the dcache, in
which case all is good, but most MIPS CPUs don't have this property).
Commit 4d46a67a3eb8 ("MIPS: Fix race condition in lazy cache flushing.")
attempted to fix that by performing the dcache flush in
flush_icache_page such that it occurs before the set_pte_at call makes
the page visible. However it has the problem that not all code that
writes to pages exposed to userland call flush_icache_page. There are
many callers of set_pte_at under mm/ and only 2 of them do call
flush_icache_page. Thus the race window between a page becoming visible
& being coherent between the icache & dcache remains open in some cases.

To illustrate some of the cases, a WARN was added to __update_cache with
this patch applied that triggered in cases where a page about to be
flushed from the dcache was not the last page provided to
flush_icache_page. That is, backtraces were obtained for cases in which
the race window is left open without this patch. The 2 standout examples
follow.

When forking a process:

[   15.271842] [<80417630>] __update_cache+0xcc/0x188
[   15.277274] [<80530394>] copy_page_range+0x56c/0x6ac
[   15.282861] [<8042936c>] copy_process.part.54+0xd40/0x17ac
[   15.289028] [<80429f80>] do_fork+0xe4/0x420
[   15.293747] [<80413808>] handle_sys+0x128/0x14c

When exec'ing an ELF binary:

[   14.445964] [<80417630>] __update_cache+0xcc/0x188
[   14.451369] [<80538d88>] move_page_tables+0x414/0x498
[   14.457075] [<8055d848>] setup_arg_pages+0x220/0x318
[   14.462685] [<805b0f38>] load_elf_binary+0x530/0x12a0
[   14.468374] [<8055ec3c>] search_binary_handler+0xbc/0x214
[   14.474444] [<8055f6c0>] do_execveat_common+0x43c/0x67c
[   14.480324] [<8055f938>] do_execve+0x38/0x44
[   14.485137] [<80413808>] handle_sys+0x128/0x14c

These code paths write into a page, call flush_dcache_page then call
set_pte_at without flush_icache_page inbetween. The end result is that
the icache can become corrupted & userland processes may execute
unexpected or invalid code, typically resulting in a reserved
instruction exception, a trap or a segfault.

Fix this race condition fully by performing any cache maintenance
required to keep the icache & dcache in sync in set_pte_at, before the
page is made valid. This has the added bonus of ensuring the cache
maintenance always happens in one location, rather than being duplicated
in flush_icache_page & update_mmu_cache. It also matches the way other
architectures solve the same problem (see arm, ia64 & powerpc).

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Reported-by: Ionela Voinescu <ionela.voinescu@imgtec.com>
Cc: Lars Persson <lars.persson@axis.com>
Fixes: 4d46a67a3eb8 ("MIPS: Fix race condition in lazy cache flushing.")
Cc: Steven J. Hill <sjhill@realitydiluted.com>
Cc: David Daney <david.daney@cavium.com>
Cc: Huacai Chen <chenhc@lemote.com>
Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Jerome Marchand <jmarchan@redhat.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Cc: stable <stable@vger.kernel.org> # v4.1+
Patchwork: https://patchwork.linux-mips.org/patch/12722/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: Handle highmem pages in __update_cache
Paul Burton [Tue, 1 Mar 2016 02:37:58 +0000 (02:37 +0000)]
MIPS: Handle highmem pages in __update_cache

[ Upstream commit f4281bba818105c7c91799abe40bc05c0dbdaa25 ]

The following patch will expose __update_cache to highmem pages. Handle
them by mapping them in for the duration of the cache maintenance, just
like in __flush_dcache_page. The code for that isn't shared because we
need the page address in __update_cache so sharing became messy. Given
that the entirity is an extra 5 lines, just duplicate it.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: Lars Persson <lars.persson@axis.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Jerome Marchand <jmarchan@redhat.com>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Cc: stable <stable@vger.kernel.org> # v4.1+
Patchwork: https://patchwork.linux-mips.org/patch/12721/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: Flush highmem pages in __flush_dcache_page
Paul Burton [Tue, 1 Mar 2016 02:37:57 +0000 (02:37 +0000)]
MIPS: Flush highmem pages in __flush_dcache_page

[ Upstream commit 234859e49a15323cf1b2331bdde7f658c4cb45fb ]

When flush_dcache_page is called on an executable page, that page is
about to be provided to userland & we can presume that the icache
contains no valid entries for its address range. However if the icache
does not fill from the dcache then we cannot presume that the pages
content has been written back as far as the memories that the dcache
will fill from (ie. L2 or further out).

This was being done for lowmem pages, but not for highmem which can lead
to icache corruption. Fix this by mapping highmem pages & flushing their
content from the dcache in __flush_dcache_page before providing the page
to userland, just as is done for lowmem pages.

Signed-off-by: Paul Burton <paul.burton@imgtec.com>
Cc: Lars Persson <lars.persson@axis.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/12720/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoRevert "powerpc/eeh: Fix crash in eeh_add_device_early() on Cell"
Guilherme G. Piccoli [Mon, 11 Apr 2016 19:17:22 +0000 (16:17 -0300)]
Revert "powerpc/eeh: Fix crash in eeh_add_device_early() on Cell"

[ Upstream commit c2078d9ef600bdbe568c89e5ddc2c6f15b7982c8 ]

This reverts commit 89a51df5ab1d38b257300b8ac940bbac3bb0eb9b.

The function eeh_add_device_early() is used to perform EEH
initialization in devices added later on the system, like in
hotplug/DLPAR scenarios. Since the commit 89a51df5ab1d ("powerpc/eeh:
Fix crash in eeh_add_device_early() on Cell") a new check was introduced
in this function - Cell has no EEH capabilities which led to kernel oops
if hotplug was performed, so checking for eeh_enabled() was introduced
to avoid the issue.

However, in architectures that EEH is present like pSeries or PowerNV,
we might reach a case in which no PCI devices are present on boot time
and so EEH is not initialized. Then, if a device is added via DLPAR for
example, eeh_add_device_early() fails because eeh_enabled() is false,
and EEH end up not being enabled at all.

This reverts the aforementioned patch since a new verification was
introduced by the commit d91dafc02f42 ("powerpc/eeh: Delay probing EEH
device during hotplug") and so the original Cell issue does not happen
anymore.

Cc: stable@vger.kernel.org # v4.1+
Reviewed-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Signed-off-by: Guilherme G. Piccoli <gpiccoli@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agopowerpc/eeh: Restore initial state in eeh_pe_reset_and_recover()
Gavin Shan [Wed, 27 Apr 2016 01:14:51 +0000 (11:14 +1000)]
powerpc/eeh: Restore initial state in eeh_pe_reset_and_recover()

[ Upstream commit 5a0cdbfd17b90a89c64a71d8aec9773ecdb20d0d ]

The function eeh_pe_reset_and_recover() is used to recover EEH
error when the passthrou device are transferred to guest and
backwards. The content in the device's config space will be lost
on PE reset issued in the middle of the recovery. The function
saves/restores it before/after the reset. However, config access
to some adapters like Broadcom BCM5719 at this point will causes
fenced PHB. The config space is always blocked and we save 0xFF's
that are restored at late point. The memory BARs are totally
corrupted, causing another EEH error upon access to one of the
memory BARs.

This restores the config space on those adapters like BCM5719
from the content saved to the EEH device when it's populated,
to resolve above issue.

Fixes: 5cfb20b9 ("powerpc/eeh: Emulate EEH recovery for VFIO devices")
Cc: stable@vger.kernel.org #v3.18+
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Reviewed-by: Russell Currey <ruscur@russell.cc>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agopowerpc/eeh: Don't report error in eeh_pe_reset_and_recover()
Gavin Shan [Wed, 27 Apr 2016 01:14:50 +0000 (11:14 +1000)]
powerpc/eeh: Don't report error in eeh_pe_reset_and_recover()

[ Upstream commit affeb0f2d3a9af419ad7ef4ac782e1540b2f7b28 ]

The function eeh_pe_reset_and_recover() is used to recover EEH
error when the passthrough device are transferred to guest and
backwards, meaning the device's driver is vfio-pci or none.
When the driver is vfio-pci that provides error_detected() error
handler only, the handler simply stops the guest and it's not
expected behaviour. On the other hand, no error handlers will
be called if we don't have a bound driver.

This ignores the error handler in eeh_pe_reset_and_recover()
that reports the error to device driver to avoid the exceptional
behaviour.

Fixes: 5cfb20b9 ("powerpc/eeh: Emulate EEH recovery for VFIO devices")
Cc: stable@vger.kernel.org #v3.18+
Signed-off-by: Gavin Shan <gwshan@linux.vnet.ibm.com>
Reviewed-by: Russell Currey <ruscur@russell.cc>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agosched/loadavg: Fix loadavg artifacts on fully idle and on fully loaded systems
Sasha Levin [Wed, 1 Jun 2016 01:11:00 +0000 (21:11 -0400)]
sched/loadavg: Fix loadavg artifacts on fully idle and on fully loaded systems

[ Upstream commit 20878232c52329f92423d27a60e48b6a6389e0dd ]

Systems show a minimal load average of 0.00, 0.01, 0.05 even when they
have no load at all.

Uptime and /proc/loadavg on all systems with kernels released during the
last five years up until kernel version 4.6-rc5, show a 5- and 15-minute
minimum loadavg of 0.01 and 0.05 respectively. This should be 0.00 on
idle systems, but the way the kernel calculates this value prevents it
from getting lower than the mentioned values.

Likewise but not as obviously noticeable, a fully loaded system with no
processes waiting, shows a maximum 1/5/15 loadavg of 1.00, 0.99, 0.95
(multiplied by number of cores).

Once the (old) load becomes 93 or higher, it mathematically can never
get lower than 93, even when the active (load) remains 0 forever.
This results in the strange 0.00, 0.01, 0.05 uptime values on idle
systems.  Note: 93/2048 = 0.0454..., which rounds up to 0.05.

It is not correct to add a 0.5 rounding (=1024/2048) here, since the
result from this function is fed back into the next iteration again,
so the result of that +0.5 rounding value then gets multiplied by
(2048-2037), and then rounded again, so there is a virtual "ghost"
load created, next to the old and active load terms.

By changing the way the internally kept value is rounded, that internal
value equivalent now can reach 0.00 on idle, and 1.00 on full load. Upon
increasing load, the internally kept load value is rounded up, when the
load is decreasing, the load value is rounded down.

The modified code was tested on nohz=off and nohz kernels. It was tested
on vanilla kernel 4.6-rc5 and on centos 7.1 kernel 3.10.0-327. It was
tested on single, dual, and octal cores system. It was tested on virtual
hosts and bare hardware. No unwanted effects have been observed, and the
problems that the patch intended to fix were indeed gone.

Tested-by: Damien Wyart <damien.wyart@free.fr>
Signed-off-by: Vik Heyndrickx <vik.heyndrickx@veribox.net>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Cc: <stable@vger.kernel.org>
Cc: Doug Smythies <dsmythies@telus.net>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Mike Galbraith <efault@gmx.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Fixes: 0f004f5a696a ("sched: Cure more NO_HZ load average woes")
Link: http://lkml.kernel.org/r/e8d32bff-d544-7748-72b5-3c86cc71f09f@veribox.net
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agortlwifi: pci: use dev_kfree_skb_irq instead of kfree_skb in rtl_pci_reset_trx_ring
Sasha Levin [Wed, 1 Jun 2016 01:07:16 +0000 (21:07 -0400)]
rtlwifi: pci: use dev_kfree_skb_irq instead of kfree_skb in rtl_pci_reset_trx_ring

[ Upstream commit cf968937d27751296920e6b82ffa89735e3a0023 ]

We can't use kfree_skb in irq disable context, because spin_lock_irqsave
make sure we are always in irq disable context, use dev_kfree_skb_irq
instead of kfree_skb is better than dev_kfree_skb_any.

This patch fix below kernel warning:
[ 7612.095528] ------------[ cut here ]------------
[ 7612.095546] WARNING: CPU: 3 PID: 4460 at kernel/softirq.c:150 __local_bh_enable_ip+0x58/0x80()
[ 7612.095550] Modules linked in: rtl8723be x86_pkg_temp_thermal btcoexist rtl_pci rtlwifi rtl8723_common
[ 7612.095567] CPU: 3 PID: 4460 Comm: ifconfig Tainted: G        W       4.4.0+ #4
[ 7612.095570] Hardware name: LENOVO 20DFA04FCD/20DFA04FCD, BIOS J5ET48WW (1.19 ) 08/27/2015
[ 7612.095574]  00000000 00000000 da37fc70 c12ce7c5 00000000 da37fca0 c104cc59 c19d4454
[ 7612.095584]  00000003 0000116c c19d4784 00000096 c10508a8 c10508a8 00000200 c1b42400
[ 7612.095594]  f29be780 da37fcb0 c104ccad 00000009 00000000 da37fcbc c10508a8 f21f08b8
[ 7612.095604] Call Trace:
[ 7612.095614]  [<c12ce7c5>] dump_stack+0x41/0x5c
[ 7612.095620]  [<c104cc59>] warn_slowpath_common+0x89/0xc0
[ 7612.095628]  [<c10508a8>] ? __local_bh_enable_ip+0x58/0x80
[ 7612.095634]  [<c10508a8>] ? __local_bh_enable_ip+0x58/0x80
[ 7612.095640]  [<c104ccad>] warn_slowpath_null+0x1d/0x20
[ 7612.095646]  [<c10508a8>] __local_bh_enable_ip+0x58/0x80
[ 7612.095653]  [<c16b7d34>] destroy_conntrack+0x64/0xa0
[ 7612.095660]  [<c16b300f>] nf_conntrack_destroy+0xf/0x20
[ 7612.095665]  [<c1677565>] skb_release_head_state+0x55/0xa0
[ 7612.095670]  [<c16775bb>] skb_release_all+0xb/0x20
[ 7612.095674]  [<c167760b>] __kfree_skb+0xb/0x60
[ 7612.095679]  [<c16776f0>] kfree_skb+0x30/0x70
[ 7612.095686]  [<f81b869d>] ? rtl_pci_reset_trx_ring+0x22d/0x370 [rtl_pci]
[ 7612.095692]  [<f81b869d>] rtl_pci_reset_trx_ring+0x22d/0x370 [rtl_pci]
[ 7612.095698]  [<f81b87f9>] rtl_pci_start+0x19/0x190 [rtl_pci]
[ 7612.095705]  [<f81970e6>] rtl_op_start+0x56/0x90 [rtlwifi]
[ 7612.095712]  [<c17e3f16>] drv_start+0x36/0xc0
[ 7612.095717]  [<c17f5ab3>] ieee80211_do_open+0x2d3/0x890
[ 7612.095725]  [<c16820fe>] ? call_netdevice_notifiers_info+0x2e/0x60
[ 7612.095730]  [<c17f60bd>] ieee80211_open+0x4d/0x50
[ 7612.095736]  [<c16891b3>] __dev_open+0xa3/0x130
[ 7612.095742]  [<c183fa53>] ? _raw_spin_unlock_bh+0x13/0x20
[ 7612.095748]  [<c1689499>] __dev_change_flags+0x89/0x140
[ 7612.095753]  [<c127c70d>] ? selinux_capable+0xd/0x10
[ 7612.095759]  [<c1689589>] dev_change_flags+0x29/0x60
[ 7612.095765]  [<c1700b93>] devinet_ioctl+0x553/0x670
[ 7612.095772]  [<c12db758>] ? _copy_to_user+0x28/0x40
[ 7612.095777]  [<c17018b5>] inet_ioctl+0x85/0xb0
[ 7612.095783]  [<c166e647>] sock_ioctl+0x67/0x260
[ 7612.095788]  [<c166e5e0>] ? sock_fasync+0x80/0x80
[ 7612.095795]  [<c115c99b>] do_vfs_ioctl+0x6b/0x550
[ 7612.095800]  [<c127c812>] ? selinux_file_ioctl+0x102/0x1e0
[ 7612.095807]  [<c10a8914>] ? timekeeping_suspend+0x294/0x320
[ 7612.095813]  [<c10a256a>] ? __hrtimer_run_queues+0x14a/0x210
[ 7612.095820]  [<c1276e24>] ? security_file_ioctl+0x34/0x50
[ 7612.095827]  [<c115cef0>] SyS_ioctl+0x70/0x80
[ 7612.095832]  [<c1001804>] do_fast_syscall_32+0x84/0x120
[ 7612.095839]  [<c183ff91>] sysenter_past_esp+0x36/0x55
[ 7612.095844] ---[ end trace 97e9c637a20e8348 ]---

Signed-off-by: Wang YanQing <udknight@gmail.com>
Cc: Stable <stable@vger.kernel.org>
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agortlwifi: Fix logic error in enter/exit power-save mode
Sasha Levin [Wed, 1 Jun 2016 01:04:01 +0000 (21:04 -0400)]
rtlwifi: Fix logic error in enter/exit power-save mode

[ Upstream commit 873ffe154ae074c46ed2d72dbd9a2a99f06f55b4 ]

In commit a269913c52ad ("rtlwifi: Rework rtl_lps_leave() and
rtl_lps_enter() to use work queue"), the tests for enter/exit
power-save mode were inverted. With this change applied, the
wifi connection becomes much more stable.

Fixes: a269913c52ad ("rtlwifi: Rework rtl_lps_leave() and rtl_lps_enter() to use work queue")
Signed-off-by: Wang YanQing <udknight@gmail.com>
CC: Stable <stable@vger.kernel.org> [3.10+]
Acked-by: Larry Finger <Larry.Finger@lwfinger.net>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agokbuild: move -Wunused-const-variable to W=1 warning level
Arnd Bergmann [Tue, 10 May 2016 21:30:01 +0000 (23:30 +0200)]
kbuild: move -Wunused-const-variable to W=1 warning level

[ Upstream commit c9c6837d39311b0cc14cdbe7c18e815ab44aefb1 ]

gcc-6 started warning by default about variables that are not
used anywhere and that are marked 'const', generating many
false positives in an allmodconfig build, e.g.:

arch/arm/mach-davinci/board-da830-evm.c:282:20: warning: 'da830_evm_emif25_pins' defined but not used [-Wunused-const-variable=]
arch/arm/plat-omap/dmtimer.c:958:34: warning: 'omap_timer_match' defined but not used [-Wunused-const-variable=]
drivers/bluetooth/hci_bcm.c:625:39: warning: 'acpi_bcm_default_gpios' defined but not used [-Wunused-const-variable=]
drivers/char/hw_random/omap-rng.c:92:18: warning: 'reg_map_omap4' defined but not used [-Wunused-const-variable=]
drivers/devfreq/exynos/exynos5_bus.c:381:32: warning: 'exynos5_busfreq_int_pm' defined but not used [-Wunused-const-variable=]
drivers/dma/mv_xor.c:1139:34: warning: 'mv_xor_dt_ids' defined but not used [-Wunused-const-variable=]

This is similar to the existing -Wunused-but-set-variable warning
that was added in an earlier release and that we disable by default
now and only enable when W=1 is set, so it makes sense to do
the same here. Once we have eliminated the majority of the
warnings for both, we can put them back into the default list.

We probably want this in backport kernels as well, to allow building
them with gcc-6 without introducing extra warnings.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Olof Johansson <olof@lixom.net>
Acked-by: Lee Jones <lee.jones@linaro.org>
Cc: stable@vger.kernel.org
Signed-off-by: Michal Marek <mmarek@suse.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoirqchip/gic-v3: Configure all interrupts as non-secure Group-1
Marc Zyngier [Fri, 6 May 2016 18:41:56 +0000 (19:41 +0100)]
irqchip/gic-v3: Configure all interrupts as non-secure Group-1

[ Upstream commit 7c9b973061b03af62734f613f6abec46c0dd4a88 ]

The GICv3 driver wrongly assumes that it runs on the non-secure
side of a secure-enabled system, while it could be on a system
with a single security state, or a GICv3 with GICD_CTLR.DS set.

Either way, it is important to configure this properly, or
interrupts will simply not be delivered on this HW.

Cc: stable@vger.kernel.org
Reported-by: Peter Maydell <peter.maydell@linaro.org>
Tested-by: Peter Maydell <peter.maydell@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoirqchip/gic: Ensure ordering between read of INTACK and shared data
Will Deacon [Tue, 26 Apr 2016 11:00:00 +0000 (12:00 +0100)]
irqchip/gic: Ensure ordering between read of INTACK and shared data

[ Upstream commit f86c4fbd930ff6fecf3d8a1c313182bd0f49f496 ]

When an IPI is generated by a CPU, the pattern looks roughly like:

  <write shared data>
  smp_wmb();
  <write to GIC to signal SGI>

On the receiving CPU we rely on the fact that, once we've taken the
interrupt, then the freshly written shared data must be visible to us.
Put another way, the CPU isn't going to speculate taking an interrupt.

Unfortunately, this assumption turns out to be broken.

Consider that CPUx wants to send an IPI to CPUy, which will cause CPUy
to read some shared_data. Before CPUx has done anything, a random
peripheral raises an IRQ to the GIC and the IRQ line on CPUy is raised.
CPUy then takes the IRQ and starts executing the entry code, heading
towards gic_handle_irq. Furthermore, let's assume that a bunch of the
previous interrupts handled by CPUy were SGIs, so the branch predictor
kicks in and speculates that irqnr will be <16 and we're likely to
head into handle_IPI. The prefetcher then grabs a speculative copy of
shared_data which contains a stale value.

Meanwhile, CPUx gets round to updating shared_data and asking the GIC
to send an SGI to CPUy. Internally, the GIC decides that the SGI is
more important than the peripheral interrupt (which hasn't yet been
ACKed) but doesn't need to do anything to CPUy, because the IRQ line
is already raised.

CPUy then reads the ACK register on the GIC, sees the SGI value which
confirms the branch prediction and we end up with a stale shared_data
value.

This patch fixes the problem by adding an smp_rmb() to the IPI entry
code in gic_handle_irq. As it turns out, the combination of a control
dependency and an ISB instruction from the EOI in the GICv3 driver is
enough to provide the ordering we need, so we add a comment there
justifying the absence of an explicit smp_rmb().

Cc: stable@vger.kernel.org
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agogcov: disable tree-loop-im to reduce stack usage
Arnd Bergmann [Mon, 25 Apr 2016 15:35:30 +0000 (17:35 +0200)]
gcov: disable tree-loop-im to reduce stack usage

[ Upstream commit c87bf431448b404a6ef5fbabd74c0e3e42157a7f ]

Enabling CONFIG_GCOV_PROFILE_ALL produces us a lot of warnings like

lib/lz4/lz4hc_compress.c: In function 'lz4_compresshcctx':
lib/lz4/lz4hc_compress.c:514:1: warning: the frame size of 1504 bytes is larger than 1024 bytes [-Wframe-larger-than=]

After some investigation, I found that this behavior started with gcc-4.9,
and opened https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69702.
A suggested workaround for it is to use the -fno-tree-loop-im
flag that turns off one of the optimization stages in gcc, so the
code runs a little slower but does not use excessive amounts
of stack.

We could make this conditional on the gcc version, but I could not
find an easy way to do this in Kbuild and the benefit would be
fairly small, given that most of the gcc version in production are
affected now.

I'm marking this for 'stable' backports because it addresses a bug
with code generation in gcc that exists in all kernel versions
with the affected gcc releases.

Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Peter Oberparleiter <oberpar@linux.vnet.ibm.com>
Cc: stable@vger.kernel.org
Signed-off-by: Michal Marek <mmarek@suse.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: KVM: Fix timer IRQ race when writing CP0_Compare
James Hogan [Fri, 22 Apr 2016 09:38:46 +0000 (10:38 +0100)]
MIPS: KVM: Fix timer IRQ race when writing CP0_Compare

[ Upstream commit b45bacd2d048f405c7760e5cc9b60dd67708734f ]

Writing CP0_Compare clears the timer interrupt pending bit
(CP0_Cause.TI), but this wasn't being done atomically. If a timer
interrupt raced with the write of the guest CP0_Compare, the timer
interrupt could end up being pending even though the new CP0_Compare is
nowhere near CP0_Count.

We were already updating the hrtimer expiry with
kvm_mips_update_hrtimer(), which used both kvm_mips_freeze_hrtimer() and
kvm_mips_resume_hrtimer(). Close the race window by expanding out
kvm_mips_update_hrtimer(), and clearing CP0_Cause.TI and setting
CP0_Compare between the freeze and resume. Since the pending timer
interrupt should not be cleared when CP0_Compare is written via the KVM
user API, an ack argument is added to distinguish the source of the
write.

Fixes: e30492bbe95a ("MIPS: KVM: Rewrite count/compare timer emulation")
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: "Radim KrÄ\8dmář" <rkrcmar@redhat.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: linux-mips@linux-mips.org
Cc: kvm@vger.kernel.org
Cc: <stable@vger.kernel.org> # 3.16.x-
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: KVM: Fix timer IRQ race when freezing timer
James Hogan [Fri, 22 Apr 2016 09:38:45 +0000 (10:38 +0100)]
MIPS: KVM: Fix timer IRQ race when freezing timer

[ Upstream commit 4355c44f063d3de4f072d796604c7f4ba4085cc3 ]

There's a particularly narrow and subtle race condition when the
software emulated guest timer is frozen which can allow a guest timer
interrupt to be missed.

This happens due to the hrtimer expiry being inexact, so very
occasionally the freeze time will be after the moment when the emulated
CP0_Count transitions to the same value as CP0_Compare (so an IRQ should
be generated), but before the moment when the hrtimer is due to expire
(so no IRQ is generated). The IRQ won't be generated when the timer is
resumed either, since the resume CP0_Count will already match CP0_Compare.

With VZ guests in particular this is far more likely to happen, since
the soft timer may be frozen frequently in order to restore the timer
state to the hardware guest timer. This happens after 5-10 hours of
guest soak testing, resulting in an overflow in guest kernel timekeeping
calculations, hanging the guest. A more focussed test case to
intentionally hit the race (with the help of a new hypcall to cause the
timer state to migrated between hardware & software) hits the condition
fairly reliably within around 30 seconds.

Instead of relying purely on the inexact hrtimer expiry to determine
whether an IRQ should be generated, read the guest CP0_Compare and
directly check whether the freeze time is before or after it. Only if
CP0_Count is on or after CP0_Compare do we check the hrtimer expiry to
determine whether the last IRQ has already been generated (which will
have pushed back the expiry by one timer period).

Fixes: e30492bbe95a ("MIPS: KVM: Rewrite count/compare timer emulation")
Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Paolo Bonzini <pbonzini@redhat.com>
Cc: "Radim KrÄ\8dmář" <rkrcmar@redhat.com>
Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: linux-mips@linux-mips.org
Cc: kvm@vger.kernel.org
Cc: <stable@vger.kernel.org> # 3.16.x-
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agocrypto: caam - fix caam_jr_alloc() ret code
Catalin Vasile [Fri, 6 May 2016 13:18:53 +0000 (16:18 +0300)]
crypto: caam - fix caam_jr_alloc() ret code

[ Upstream commit e930c765ca5c6b039cd22ebfb4504ea7b5dab43d ]

caam_jr_alloc() used to return NULL if a JR device could not be
allocated for a session. In turn, every user of this function used
IS_ERR() function to verify if anything went wrong, which does NOT look
for NULL values. This made the kernel crash if the sanity check failed,
because the driver continued to think it had allocated a valid JR dev
instance to the session and at some point it tries to do a caam_jr_free()
on a NULL JR dev pointer.
This patch is a fix for this issue.

Cc: <stable@vger.kernel.org>
Signed-off-by: Catalin Vasile <cata.vasile@nxp.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: serial: quatech2: fix use-after-free in probe error path
Johan Hovold [Sun, 8 May 2016 18:08:02 +0000 (20:08 +0200)]
USB: serial: quatech2: fix use-after-free in probe error path

[ Upstream commit 028c49f5e02a257c94129cd815f7c8485f51d4ef ]

The interface read URB is submitted in attach, but was only unlinked by
the driver at disconnect.

In case of a late probe error (e.g. due to failed minor allocation),
disconnect is never called and we would end up with active URBs for an
unbound interface. This in turn could lead to deallocated memory being
dereferenced in the completion callback.

Fixes: f7a33e608d9a ("USB: serial: add quatech2 usb to serial driver")
Cc: stable <stable@vger.kernel.org> # v3.5: 40d04738491d
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: serial: mxuport: fix use-after-free in probe error path
Johan Hovold [Sun, 8 May 2016 18:08:01 +0000 (20:08 +0200)]
USB: serial: mxuport: fix use-after-free in probe error path

[ Upstream commit 9e45284984096314994777f27e1446dfbfd2f0d7 ]

The interface read and event URBs are submitted in attach, but were
never explicitly unlinked by the driver. Instead the URBs would have
been killed by usb-serial core on disconnect.

In case of a late probe error (e.g. due to failed minor allocation),
disconnect is never called and we could end up with active URBs for an
unbound interface. This in turn could lead to deallocated memory being
dereferenced in the completion callbacks.

Fixes: ee467a1f2066 ("USB: serial: add Moxa UPORT 12XX/14XX/16XX
driver")
Cc: stable <stable@vger.kernel.org> # v3.14
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: serial: keyspan: fix use-after-free in probe error path
Johan Hovold [Sun, 8 May 2016 18:07:58 +0000 (20:07 +0200)]
USB: serial: keyspan: fix use-after-free in probe error path

[ Upstream commit 35be1a71d70775e7bd7e45fa6d2897342ff4c9d2 ]

The interface instat and indat URBs were submitted in attach, but never
unlinked in release before deallocating the corresponding transfer
buffers.

In the case of a late probe error (e.g. due to failed minor allocation),
disconnect would not have been called before release, causing the
buffers to be freed while the URBs are still in use. We'd also end up
with active URBs for an unbound interface.

Fixes: f9c99bb8b3a1 ("USB: usb-serial: replace shutdown with disconnect,
release")
Cc: stable <stable@vger.kernel.org> # v2.6.31
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: serial: io_edgeport: fix memory leaks in probe error path
Johan Hovold [Sun, 8 May 2016 18:07:57 +0000 (20:07 +0200)]
USB: serial: io_edgeport: fix memory leaks in probe error path

[ Upstream commit c8d62957d450cc1a22ce3242908709fe367ddc8e ]

URBs and buffers allocated in attach for Epic devices would never be
deallocated in case of a later probe error (e.g. failure to allocate
minor numbers) as disconnect is then never called.

Fix by moving deallocation to release and making sure that the
URBs are first unlinked.

Fixes: f9c99bb8b3a1 ("USB: usb-serial: replace shutdown with disconnect,
release")
Cc: stable <stable@vger.kernel.org> # v2.6.31
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: serial: io_edgeport: fix memory leaks in attach error path
Johan Hovold [Sun, 8 May 2016 18:07:56 +0000 (20:07 +0200)]
USB: serial: io_edgeport: fix memory leaks in attach error path

[ Upstream commit c5c0c55598cefc826d6cfb0a417eeaee3631715c ]

Private data, URBs and buffers allocated for Epic devices during
attach were never released on errors (e.g. missing endpoints).

Fixes: 6e8cf7751f9f ("USB: add EPIC support to the io_edgeport driver")
Cc: stable <stable@vger.kernel.org> # v2.6.21
Signed-off-by: Johan Hovold <johan@kernel.org>
Acked-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agomfd: omap-usb-tll: Fix scheduling while atomic BUG
Roger Quadros [Mon, 9 May 2016 08:28:37 +0000 (11:28 +0300)]
mfd: omap-usb-tll: Fix scheduling while atomic BUG

[ Upstream commit b49b927f16acee626c56a1af4ab4cb062f75b5df ]

We shouldn't be calling clk_prepare_enable()/clk_prepare_disable()
in an atomic context.

Fixes the following issue:

[    5.830970] ehci-omap: OMAP-EHCI Host Controller driver
[    5.830974] driver_register 'ehci-omap'
[    5.895849] driver_register 'wl1271_sdio'
[    5.896870] BUG: scheduling while atomic: udevd/994/0x00000002
[    5.896876] 4 locks held by udevd/994:
[    5.896904]  #0:  (&dev->mutex){......}, at: [<c049597c>] __driver_attach+0x60/0xac
[    5.896923]  #1:  (&dev->mutex){......}, at: [<c049598c>] __driver_attach+0x70/0xac
[    5.896946]  #2:  (tll_lock){+.+...}, at: [<c04c2630>] omap_tll_enable+0x2c/0xd0
[    5.896966]  #3:  (prepare_lock){+.+...}, at: [<c05ce9c8>] clk_prepare_lock+0x48/0xe0
[    5.897042] Modules linked in: wlcore_sdio(+) ehci_omap(+) dwc3_omap snd_soc_ts3a225e leds_is31fl319x bq27xxx_battery_i2c tsc2007 bq27xxx_battery bq2429x_charger ina2xx tca8418_keypad as5013 leds_tca6507 twl6040_vibra gpio_twl6040 bmp085_i2c(+) palmas_gpadc usb3503 palmas_pwrbutton bmg160_i2c(+) bmp085 bma150(+) bmg160_core bmp280 input_polldev snd_soc_omap_mcbsp snd_soc_omap_mcpdm snd_soc_omap snd_pcm_dmaengine
[    5.897048] Preemption disabled at:[<  (null)>]   (null)
[    5.897051]
[    5.897059] CPU: 0 PID: 994 Comm: udevd Not tainted 4.6.0-rc5-letux+ #233
[    5.897062] Hardware name: Generic OMAP5 (Flattened Device Tree)
[    5.897076] [<c010e714>] (unwind_backtrace) from [<c010af34>] (show_stack+0x10/0x14)
[    5.897087] [<c010af34>] (show_stack) from [<c040aa7c>] (dump_stack+0x88/0xc0)
[    5.897099] [<c040aa7c>] (dump_stack) from [<c020c558>] (__schedule_bug+0xac/0xd0)
[    5.897111] [<c020c558>] (__schedule_bug) from [<c06f3d44>] (__schedule+0x88/0x7e4)
[    5.897120] [<c06f3d44>] (__schedule) from [<c06f46d8>] (schedule+0x9c/0xc0)
[    5.897129] [<c06f46d8>] (schedule) from [<c06f4904>] (schedule_preempt_disabled+0x14/0x20)
[    5.897140] [<c06f4904>] (schedule_preempt_disabled) from [<c06f64e4>] (mutex_lock_nested+0x258/0x43c)
[    5.897150] [<c06f64e4>] (mutex_lock_nested) from [<c05ce9c8>] (clk_prepare_lock+0x48/0xe0)
[    5.897160] [<c05ce9c8>] (clk_prepare_lock) from [<c05d0e7c>] (clk_prepare+0x10/0x28)
[    5.897169] [<c05d0e7c>] (clk_prepare) from [<c04c2668>] (omap_tll_enable+0x64/0xd0)
[    5.897180] [<c04c2668>] (omap_tll_enable) from [<c04c1728>] (usbhs_runtime_resume+0x18/0x17c)
[    5.897192] [<c04c1728>] (usbhs_runtime_resume) from [<c049d404>] (pm_generic_runtime_resume+0x2c/0x40)
[    5.897202] [<c049d404>] (pm_generic_runtime_resume) from [<c049f180>] (__rpm_callback+0x38/0x68)
[    5.897210] [<c049f180>] (__rpm_callback) from [<c049f220>] (rpm_callback+0x70/0x88)
[    5.897218] [<c049f220>] (rpm_callback) from [<c04a0a00>] (rpm_resume+0x4ec/0x7ec)
[    5.897227] [<c04a0a00>] (rpm_resume) from [<c04a0f48>] (__pm_runtime_resume+0x4c/0x64)
[    5.897236] [<c04a0f48>] (__pm_runtime_resume) from [<c04958dc>] (driver_probe_device+0x30/0x70)
[    5.897246] [<c04958dc>] (driver_probe_device) from [<c04959a4>] (__driver_attach+0x88/0xac)
[    5.897256] [<c04959a4>] (__driver_attach) from [<c04940f8>] (bus_for_each_dev+0x50/0x84)
[    5.897267] [<c04940f8>] (bus_for_each_dev) from [<c0494e40>] (bus_add_driver+0xcc/0x1e4)
[    5.897276] [<c0494e40>] (bus_add_driver) from [<c0496914>] (driver_register+0xac/0xf4)
[    5.897286] [<c0496914>] (driver_register) from [<c01018e0>] (do_one_initcall+0x100/0x1b8)
[    5.897296] [<c01018e0>] (do_one_initcall) from [<c01c7a54>] (do_init_module+0x58/0x1c0)
[    5.897304] [<c01c7a54>] (do_init_module) from [<c01c8a3c>] (SyS_finit_module+0x88/0x90)
[    5.897313] [<c01c8a3c>] (SyS_finit_module) from [<c0107120>] (ret_fast_syscall+0x0/0x1c)
[    5.912697] ------------[ cut here ]------------
[    5.912711] WARNING: CPU: 0 PID: 994 at kernel/sched/core.c:2996 _raw_spin_unlock+0x28/0x58
[    5.912717] DEBUG_LOCKS_WARN_ON(val > preempt_count())

Cc: <stable@vger.kernel.org>
Reported-by: H. Nikolaus Schaller <hns@goldelico.com>
Tested-by: H. Nikolaus Schaller <hns@goldelico.com>
Signed-off-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS64: R6: R2 emulation bugfix
Leonid Yegoshin [Tue, 28 Apr 2015 19:53:35 +0000 (12:53 -0700)]
MIPS64: R6: R2 emulation bugfix

[ Upstream commit 41fa29e4d8cf4150568a0fe9bb4d62229f9caed5 ]

Error recovery pointers for fixups was improperly set as ".word"
which is unsuitable for MIPS64.

Replaced by STR(PTR)

[ralf@linux-mips.org: Apply changes as requested in the review process.]

Signed-off-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Reviewed-by: James Hogan <james.hogan@imgtec.com>
Reviewed-by: Markos Chandras <markos.chandras@imgtec.com>
Fixes: b0a668fb2038 ("MIPS: kernel: mips-r2-to-r6-emul: Add R2 emulator for MIPS R6")
Cc: macro@linux-mips.org
Cc: linux-mips@linux-mips.org
Cc: linux-kernel@vger.kernel.org
Cc: <stable@vger.kernel.org> # 4.0+
Patchwork: https://patchwork.linux-mips.org/patch/9911/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: Avoid using unwind_stack() with usermode
James Hogan [Fri, 4 Dec 2015 22:25:02 +0000 (22:25 +0000)]
MIPS: Avoid using unwind_stack() with usermode

[ Upstream commit 81a76d7119f63c359750e4adeff922a31ad1135f ]

When showing backtraces in response to traps, for example crashes and
address errors (usually unaligned accesses) when they are set in debugfs
to be reported, unwind_stack will be used if the PC was in the kernel
text address range. However since EVA it is possible for user and kernel
address ranges to overlap, and even without EVA userland can still
trigger an address error by jumping to a KSeg0 address.

Adjust the check to also ensure that it was running in kernel mode. I
don't believe any harm can come of this problem, since unwind_stack() is
sufficiently defensive, however it is only meant for unwinding kernel
code, so to be correct it should use the raw backtracing instead.

Signed-off-by: James Hogan <james.hogan@imgtec.com>
Reviewed-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Cc: linux-mips@linux-mips.org
Cc: <stable@vger.kernel.org> # 3.15+
Patchwork: https://patchwork.linux-mips.org/patch/11701/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
(cherry picked from commit d2941a975ac745c607dfb590e92bb30bc352dad9)
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: Don't unwind to user mode with EVA
James Hogan [Fri, 4 Dec 2015 22:25:01 +0000 (22:25 +0000)]
MIPS: Don't unwind to user mode with EVA

[ Upstream commit a816b306c62195b7c43c92cb13330821a96bdc27 ]

When unwinding through IRQs and exceptions, the unwinding only continues
if the PC is a kernel text address, however since EVA it is possible for
user and kernel address ranges to overlap, potentially allowing
unwinding to continue to user mode if the user PC happens to be in the
kernel text address range.

Adjust the check to also ensure that the register state from before the
exception is actually running in kernel mode, i.e. !user_mode(regs).

I don't believe any harm can come of this problem, since the PC is only
output, the stack pointer is checked to ensure it resides within the
task's stack page before it is dereferenced in search of the return
address, and the return address register is similarly only output (if
the PC is in a leaf function or the beginning of a non-leaf function).

However unwind_stack() is only meant for unwinding kernel code, so to be
correct the unwind should stop there.

Signed-off-by: James Hogan <james.hogan@imgtec.com>
Reviewed-by: Leonid Yegoshin <Leonid.Yegoshin@imgtec.com>
Cc: linux-mips@linux-mips.org
Cc: <stable@vger.kernel.org> # 3.15+
Patchwork: https://patchwork.linux-mips.org/patch/11700/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: Fix siginfo.h to use strict posix types
James Hogan [Mon, 8 Feb 2016 18:43:49 +0000 (18:43 +0000)]
MIPS: Fix siginfo.h to use strict posix types

[ Upstream commit 5daebc477da4dfeb31ae193d83084def58fd2697 ]

Commit 85efde6f4e0d ("make exported headers use strict posix types")
changed the asm-generic siginfo.h to use the __kernel_* types, and
commit 3a471cbc081b ("remove __KERNEL_STRICT_NAMES") make the internal
types accessible only to the kernel, but the MIPS implementation hasn't
been updated to match.

Switch to proper types now so that the exported asm/siginfo.h won't
produce quite so many compiler errors when included alone by a user
program.

Signed-off-by: James Hogan <james.hogan@imgtec.com>
Cc: Christopher Ferris <cferris@google.com>
Cc: linux-mips@linux-mips.org
Cc: <stable@vger.kernel.org> # 2.6.30-
Cc: linux-kernel@vger.kernel.org
Patchwork: https://patchwork.linux-mips.org/patch/12477/
Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agocan: fix handling of unmodifiable configuration options
Oliver Hartkopp [Mon, 21 Mar 2016 19:18:21 +0000 (20:18 +0100)]
can: fix handling of unmodifiable configuration options

[ Upstream commit bb208f144cf3f59d8f89a09a80efd04389718907 ]

As described in 'can: m_can: tag current CAN FD controllers as non-ISO'
(6cfda7fbebe) it is possible to define fixed configuration options by
setting the according bit in 'ctrlmode' and clear it in 'ctrlmode_supported'.
This leads to the incovenience that the fixed configuration bits can not be
passed by netlink even when they have the correct values (e.g. non-ISO, FD).

This patch fixes that issue and not only allows fixed set bit values to be set
again but now requires(!) to provide these fixed values at configuration time.
A valid CAN FD configuration consists of a nominal/arbitration bittiming, a
data bittiming and a control mode with CAN_CTRLMODE_FD set - which is now
enforced by a new can_validate() function. This fix additionally removed the
inconsistency that was prohibiting the support of 'CANFD-only' controller
drivers, like the RCar CAN FD.

For this reason a new helper can_set_static_ctrlmode() has been introduced to
provide a proper interface to handle static enabled CAN controller options.

Reported-by: Ramesh Shanmugasundaram <ramesh.shanmugasundaram@bp.renesas.com>
Signed-off-by: Oliver Hartkopp <socketcan@hartkopp.net>
Reviewed-by: Ramesh Shanmugasundaram <ramesh.shanmugasundaram@bp.renesas.com>
Cc: <stable@vger.kernel.org> # >= 3.18
Signed-off-by: Marc Kleine-Budde <mkl@pengutronix.de>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoarm64: Ensure pmd_present() returns false after pmd_mknotpresent()
Catalin Marinas [Thu, 5 May 2016 09:44:02 +0000 (10:44 +0100)]
arm64: Ensure pmd_present() returns false after pmd_mknotpresent()

[ Upstream commit 5bb1cc0ff9a6b68871970737e6c4c16919928d8b ]

Currently, pmd_present() only checks for a non-zero value, returning
true even after pmd_mknotpresent() (which only clears the type bits).
This patch converts pmd_present() to using pte_present(), similar to the
other pmd_*() checks. As a side effect, it will return true for
PROT_NONE mappings, though they are not yet used by the kernel with
transparent huge pages.

For consistency, also change pmd_mknotpresent() to only clear the
PMD_SECT_VALID bit, even though the PMD_TABLE_BIT is already 0 for block
mappings (no functional change). The unused PMD_SECT_PROT_NONE
definition is removed as transparent huge pages use the pte page prot
values.

Fixes: 9c7e535fcc17 ("arm64: mm: Route pmd thp functions through pte equivalents")
Cc: <stable@vger.kernel.org> # 3.15+
Reviewed-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Catalin Marinas <catalin.marinas@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoext4: silence UBSAN in ext4_mb_init()
Nicolai Stange [Thu, 5 May 2016 23:46:19 +0000 (19:46 -0400)]
ext4: silence UBSAN in ext4_mb_init()

[ Upstream commit 935244cd54b86ca46e69bc6604d2adfb1aec2d42 ]

Currently, in ext4_mb_init(), there's a loop like the following:

  do {
    ...
    offset += 1 << (sb->s_blocksize_bits - i);
    i++;
  } while (i <= sb->s_blocksize_bits + 1);

Note that the updated offset is used in the loop's next iteration only.

However, at the last iteration, that is at i == sb->s_blocksize_bits + 1,
the shift count becomes equal to (unsigned)-1 > 31 (c.f. C99 6.5.7(3))
and UBSAN reports

  UBSAN: Undefined behaviour in fs/ext4/mballoc.c:2621:15
  shift exponent 4294967295 is too large for 32-bit type 'int'
  [...]
  Call Trace:
   [<ffffffff818c4d25>] dump_stack+0xbc/0x117
   [<ffffffff818c4c69>] ? _atomic_dec_and_lock+0x169/0x169
   [<ffffffff819411ab>] ubsan_epilogue+0xd/0x4e
   [<ffffffff81941cac>] __ubsan_handle_shift_out_of_bounds+0x1fb/0x254
   [<ffffffff81941ab1>] ? __ubsan_handle_load_invalid_value+0x158/0x158
   [<ffffffff814b6dc1>] ? kmem_cache_alloc+0x101/0x390
   [<ffffffff816fc13b>] ? ext4_mb_init+0x13b/0xfd0
   [<ffffffff814293c7>] ? create_cache+0x57/0x1f0
   [<ffffffff8142948a>] ? create_cache+0x11a/0x1f0
   [<ffffffff821c2168>] ? mutex_lock+0x38/0x60
   [<ffffffff821c23ab>] ? mutex_unlock+0x1b/0x50
   [<ffffffff814c26ab>] ? put_online_mems+0x5b/0xc0
   [<ffffffff81429677>] ? kmem_cache_create+0x117/0x2c0
   [<ffffffff816fcc49>] ext4_mb_init+0xc49/0xfd0
   [...]

Observe that the mentioned shift exponent, 4294967295, equals (unsigned)-1.

Unless compilers start to do some fancy transformations (which at least
GCC 6.0.0 doesn't currently do), the issue is of cosmetic nature only: the
such calculated value of offset is never used again.

Silence UBSAN by introducing another variable, offset_incr, holding the
next increment to apply to offset and adjust that one by right shifting it
by one position per loop iteration.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=114701
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=112161

Cc: stable@vger.kernel.org
Signed-off-by: Nicolai Stange <nicstange@gmail.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoext4: address UBSAN warning in mb_find_order_for_block()
Nicolai Stange [Thu, 5 May 2016 21:38:03 +0000 (17:38 -0400)]
ext4: address UBSAN warning in mb_find_order_for_block()

[ Upstream commit b5cb316cdf3a3f5f6125412b0f6065185240cfdc ]

Currently, in mb_find_order_for_block(), there's a loop like the following:

  while (order <= e4b->bd_blkbits + 1) {
    ...
    bb += 1 << (e4b->bd_blkbits - order);
  }

Note that the updated bb is used in the loop's next iteration only.

However, at the last iteration, that is at order == e4b->bd_blkbits + 1,
the shift count becomes negative (c.f. C99 6.5.7(3)) and UBSAN reports

  UBSAN: Undefined behaviour in fs/ext4/mballoc.c:1281:11
  shift exponent -1 is negative
  [...]
  Call Trace:
   [<ffffffff818c4d35>] dump_stack+0xbc/0x117
   [<ffffffff818c4c79>] ? _atomic_dec_and_lock+0x169/0x169
   [<ffffffff819411bb>] ubsan_epilogue+0xd/0x4e
   [<ffffffff81941cbc>] __ubsan_handle_shift_out_of_bounds+0x1fb/0x254
   [<ffffffff81941ac1>] ? __ubsan_handle_load_invalid_value+0x158/0x158
   [<ffffffff816e93a0>] ? ext4_mb_generate_from_pa+0x590/0x590
   [<ffffffff816502c8>] ? ext4_read_block_bitmap_nowait+0x598/0xe80
   [<ffffffff816e7b7e>] mb_find_order_for_block+0x1ce/0x240
   [...]

Unless compilers start to do some fancy transformations (which at least
GCC 6.0.0 doesn't currently do), the issue is of cosmetic nature only: the
such calculated value of bb is never used again.

Silence UBSAN by introducing another variable, bb_incr, holding the next
increment to apply to bb and adjust that one by right shifting it by one
position per loop iteration.

Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=114701
Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=112161

Cc: stable@vger.kernel.org
Signed-off-by: Nicolai Stange <nicstange@gmail.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoext4: fix oops on corrupted filesystem
Jan Kara [Thu, 5 May 2016 15:10:15 +0000 (11:10 -0400)]
ext4: fix oops on corrupted filesystem

[ Upstream commit 74177f55b70e2f2be770dd28684dd6d17106a4ba ]

When filesystem is corrupted in the right way, it can happen
ext4_mark_iloc_dirty() in ext4_orphan_add() returns error and we
subsequently remove inode from the in-memory orphan list. However this
deletion is done with list_del(&EXT4_I(inode)->i_orphan) and thus we
leave i_orphan list_head with a stale content. Later we can look at this
content causing list corruption, oops, or other issues. The reported
trace looked like:

WARNING: CPU: 0 PID: 46 at lib/list_debug.c:53 __list_del_entry+0x6b/0x100()
list_del corruption, 0000000061c1d6e0->next is LIST_POISON1
0000000000100100)
CPU: 0 PID: 46 Comm: ext4.exe Not tainted 4.1.0-rc4+ #250
Stack:
 60462947 62219960 602ede24 62219960
 602ede24 603ca293 622198f0 602f02eb
 62219950 6002c12c 62219900 601b4d6b
Call Trace:
 [<6005769c>] ? vprintk_emit+0x2dc/0x5c0
 [<602ede24>] ? printk+0x0/0x94
 [<600190bc>] show_stack+0xdc/0x1a0
 [<602ede24>] ? printk+0x0/0x94
 [<602ede24>] ? printk+0x0/0x94
 [<602f02eb>] dump_stack+0x2a/0x2c
 [<6002c12c>] warn_slowpath_common+0x9c/0xf0
 [<601b4d6b>] ? __list_del_entry+0x6b/0x100
 [<6002c254>] warn_slowpath_fmt+0x94/0xa0
 [<602f4d09>] ? __mutex_lock_slowpath+0x239/0x3a0
 [<6002c1c0>] ? warn_slowpath_fmt+0x0/0xa0
 [<60023ebf>] ? set_signals+0x3f/0x50
 [<600a205a>] ? kmem_cache_free+0x10a/0x180
 [<602f4e88>] ? mutex_lock+0x18/0x30
 [<601b4d6b>] __list_del_entry+0x6b/0x100
 [<601177ec>] ext4_orphan_del+0x22c/0x2f0
 [<6012f27c>] ? __ext4_journal_start_sb+0x2c/0xa0
 [<6010b973>] ? ext4_truncate+0x383/0x390
 [<6010bc8b>] ext4_write_begin+0x30b/0x4b0
 [<6001bb50>] ? copy_from_user+0x0/0xb0
 [<601aa840>] ? iov_iter_fault_in_readable+0xa0/0xc0
 [<60072c4f>] generic_perform_write+0xaf/0x1e0
 [<600c4166>] ? file_update_time+0x46/0x110
 [<60072f0f>] __generic_file_write_iter+0x18f/0x1b0
 [<6010030f>] ext4_file_write_iter+0x15f/0x470
 [<60094e10>] ? unlink_file_vma+0x0/0x70
 [<6009b180>] ? unlink_anon_vmas+0x0/0x260
 [<6008f169>] ? free_pgtables+0xb9/0x100
 [<600a6030>] __vfs_write+0xb0/0x130
 [<600a61d5>] vfs_write+0xa5/0x170
 [<600a63d6>] SyS_write+0x56/0xe0
 [<6029fcb0>] ? __libc_waitpid+0x0/0xa0
 [<6001b698>] handle_syscall+0x68/0x90
 [<6002633d>] userspace+0x4fd/0x600
 [<6002274f>] ? save_registers+0x1f/0x40
 [<60028bd7>] ? arch_prctl+0x177/0x1b0
 [<60017bd5>] fork_handler+0x85/0x90

Fix the problem by using list_del_init() as we always should with
i_orphan list.

CC: stable@vger.kernel.org
Reported-by: Vegard Nossum <vegard.nossum@oracle.com>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: serial: cp210x: fix hardware flow-control disable
Konstantin Shkolnyy [Wed, 4 May 2016 21:56:52 +0000 (16:56 -0500)]
USB: serial: cp210x: fix hardware flow-control disable

[ Upstream commit a377f9e906af4df9071ba8ddba60188cb4013d93 ]

A bug in the CRTSCTS handling caused RTS to alternate between

CRTSCTS=0 => "RTS is transmit active signal" and
CRTSCTS=1 => "RTS is used for receive flow control"

instead of

CRTSCTS=0 => "RTS is statically active" and
CRTSCTS=1 => "RTS is used for receive flow control"

This only happened after first having enabled CRTSCTS.

Signed-off-by: Konstantin Shkolnyy <konstantin.shkolnyy@gmail.com>
Fixes: 39a66b8d22a3 ("[PATCH] USB: CP2101 Add support for flow control")
Cc: stable <stable@vger.kernel.org>
[johan: reword commit message ]
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: cp210x: relocate private data from USB interface to port
Konstantin Shkolnyy [Wed, 28 Oct 2015 21:02:34 +0000 (16:02 -0500)]
USB: cp210x: relocate private data from USB interface to port

[ Upstream commit e2ae67a3b55188b0342522d8139acf013feb2a69 ]

This change is preparation for implementing a cp2108 bug workaround.
The workaround requires storing some private data. Right now the data is
attached to the USB interface and allocated in the attach() callback.
The bug detection requires USB I/O which is done easier from port_probe()
callback rather than attach(). Since the USB access functions take port
as a parameter, and since the private data is used exclusively by these
functions, it can be allocated in port_probe(). Also, all cp210x devices
have exactly 1 port per USB iterface, so moving private data from the USB
interface to port is trivial.

Signed-off-by: Konstantin Shkolnyy <konstantin.shkolnyy@gmail.com>
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoACPI / osi: Fix an issue that acpi_osi=!* cannot disable ACPICA internal strings
Lv Zheng [Tue, 3 May 2016 08:48:20 +0000 (16:48 +0800)]
ACPI / osi: Fix an issue that acpi_osi=!* cannot disable ACPICA internal strings

[ Upstream commit 30c9bb0d7603e7b3f4d6a0ea231e1cddae020c32 ]

The order of the _OSI related functionalities is as follows:

  acpi_blacklisted()
    acpi_dmi_osi_linux()
      acpi_osi_setup()
    acpi_osi_setup()
      acpi_update_interfaces() if "!*"
      <<<<<<<<<<<<<<<<<<<<<<<<
  parse_args()
    __setup("acpi_osi=")
      acpi_osi_setup_linux()
        acpi_update_interfaces() if "!*"
        <<<<<<<<<<<<<<<<<<<<<<<<
  acpi_early_init()
    acpi_initialize_subsystem()
      acpi_ut_initialize_interfaces()
      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  acpi_bus_init()
    acpi_os_initialize1()
      acpi_install_interface_handler(acpi_osi_handler)
      acpi_osi_setup_late()
        acpi_update_interfaces() for "!"
        >>>>>>>>>>>>>>>>>>>>>>>>
  acpi_osi_handler()

Since acpi_osi_setup_linux() can override acpi_dmi_osi_linux(), the command
line setting can override the DMI detection. That's why acpi_blacklisted()
is put before __setup("acpi_osi=").

Then we can notice the following wrong invocation order. There are
acpi_update_interfaces() (marked by <<<<) calls invoked before
acpi_ut_initialize_interfaces() (marked by ^^^^). This makes it impossible
to use acpi_osi=!* correctly from OSI DMI table or from the command line.
The use of acpi_osi=!* is meant to disable both ACPICA
(acpi_gbl_supported_interfaces) and Linux specific strings
(osi_setup_entries) while the ACPICA part should have stopped working
because of the order issue.

This patch fixes this issue by moving acpi_update_interfaces() to where
it is invoked for acpi_osi=! (marked by >>>>) as this is ensured to be
invoked after acpi_ut_initialize_interfaces() (marked by ^^^^). Linux
specific strings are still handled in the original place in order to make
the following command line working: acpi_osi=!* acpi_osi="Module Device".

Note that since acpi_osi=!* is meant to further disable linux specific
string comparing to the acpi_osi=!, there is no such use case in our bug
fixing work and hence there is no one using acpi_osi=!* either from the
command line or from the DMI quirks, this issue is just a theoretical
issue.

Fixes: 741d81280ad2 (ACPI: Add facility to remove all _OSI strings)
Cc: 3.12+ <stable@vger.kernel.org> # 3.12+
Tested-by: Lukas Wunner <lukas@wunner.de>
Tested-by: Chen Yu <yu.c.chen@intel.com>
Signed-off-by: Lv Zheng <lv.zheng@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: serial: option: add even more ZTE device ids
Lei Liu [Wed, 4 May 2016 08:34:22 +0000 (16:34 +0800)]
USB: serial: option: add even more ZTE device ids

[ Upstream commit 74d2a91aec97ab832790c9398d320413ad185321 ]

Add even more ZTE device ids.

Signed-off-by: lei liu <liu.lei78@zte.com.cn>
Cc: stable <stable@vger.kernel.org>
[johan: rebase and replace commit message ]
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: serial: option: add more ZTE device ids
lei liu [Tue, 3 May 2016 21:44:19 +0000 (14:44 -0700)]
USB: serial: option: add more ZTE device ids

[ Upstream commit f0d09463c59c2d764a6c6d492cbe6d2c77f27153 ]

More ZTE device ids.

Signed-off-by: lei liu <liu.lei78@zte.com.cn>
Cc: stable <stable@vger.kernel.org>
[properly sort them - gregkh]
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agomcb: Fixed bar number assignment for the gdd
Andreas Werner [Tue, 3 May 2016 10:42:00 +0000 (12:42 +0200)]
mcb: Fixed bar number assignment for the gdd

[ Upstream commit f75564d343010b025301d9548f2304f48eb25f01 ]

The bar number is found in reg2 within the gdd. Therefore
we need to change the assigment from reg1 to reg2 which
is the correct location.

Signed-off-by: Andreas Werner <andreas.werner@men.de>
Fixes: '3764e82e5' drivers: Introduce MEN Chameleon Bus
Cc: stable@vger.kernel.org # v3.15+
Signed-off-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agousb: misc: usbtest: fix pattern tests for scatterlists.
Mathias Nyman [Mon, 2 May 2016 08:39:03 +0000 (11:39 +0300)]
usb: misc: usbtest: fix pattern tests for scatterlists.

[ Upstream commit cdc77c82a8286b1181b81b6e5ef60c8e83ded7bc ]

The current implemenentation restart the sent pattern for each entry in
the sg list. The receiving end expects a continuous pattern, and test
will fail unless scatterilst entries happen to be aligned with the
pattern

Fix this by calculating the pattern byte based on total sent size
instead of just the current sg entry.

Signed-off-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Fixes: 8b5249019352 ("[PATCH] USB: usbtest: scatterlist OUT data pattern testing")
Cc: <stable@vger.kernel.org> # v2.6.18+
Acked-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agousb: misc: usbtest: format the data pattern according to max packet size
Alan Stern [Tue, 1 Sep 2015 01:48:01 +0000 (09:48 +0800)]
usb: misc: usbtest: format the data pattern according to max packet size

[ Upstream commit b9a6e8e1001e28fecbd74c073f5503dac2790563 ]

With this change, the host and gadget doesn't need to agree with transfer
length for comparing the data, since they doesn't know each other's
transfer size, but know max packet size.

Signed-off-by: Peter Chen <peter.chen@freescale.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
(Fixed the 'line over 80 characters warning' by Peter Chen)
Tested-by: Peter Chen <peter.chen@freescale.com>
Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: leave LPM alone if possible when binding/unbinding interface drivers
Alan Stern [Fri, 29 Apr 2016 19:25:17 +0000 (15:25 -0400)]
USB: leave LPM alone if possible when binding/unbinding interface drivers

[ Upstream commit 6fb650d43da3e7054984dc548eaa88765a94d49f ]

When a USB driver is bound to an interface (either through probing or
by claiming it) or is unbound from an interface, the USB core always
disables Link Power Management during the transition and then
re-enables it afterward.  The reason is because the driver might want
to prevent hub-initiated link power transitions, in which case the HCD
would have to recalculate the various LPM parameters.  This
recalculation takes place when LPM is re-enabled and the new
parameters are sent to the device and its parent hub.

However, if the driver does not want to prevent hub-initiated link
power transitions then none of this work is necessary.  The parameters
don't need to be recalculated, and LPM doesn't need to be disabled and
re-enabled.

It turns out that disabling and enabling LPM can be time-consuming,
enough so that it interferes with user programs that want to claim and
release interfaces rapidly via usbfs.  Since the usbfs kernel driver
doesn't set the disable_hub_initiated_lpm flag, we can speed things up
and get the user programs to work by leaving LPM alone whenever the
flag isn't set.

And while we're improving the way disable_hub_initiated_lpm gets used,
let's also fix its kerneldoc.

Signed-off-by: Alan Stern <stern@rowland.harvard.edu>
Tested-by: Matthew Giassa <matthew@giassa.net>
CC: Mathias Nyman <mathias.nyman@intel.com>
CC: <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoUSB: serial: option: add support for Cinterion PH8 and AHxx
Schemmel Hans-Christoph [Fri, 29 Apr 2016 08:51:06 +0000 (08:51 +0000)]
USB: serial: option: add support for Cinterion PH8 and AHxx

[ Upstream commit 444f94e9e625f6ec6bbe2cb232a6451c637f35a3 ]

Added support for Gemalto's Cinterion PH8 and AHxx products
with 2 RmNet Interfaces and products with 1 RmNet + 1 USB Audio interface.

In addition some minor renaming and formatting.

Signed-off-by: Hans-Christoph Schemmel <hans-christoph.schemmel@gemalto.com>
[johan: sort current entries and trim trailing whitespace ]
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Johan Hovold <johan@kernel.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agothunderbolt: Fix double free of drom buffer
Andreas Noever [Sun, 10 Apr 2016 10:48:27 +0000 (12:48 +0200)]
thunderbolt: Fix double free of drom buffer

[ Upstream commit 2ffa9a5d76a75abbc1f95c17959fced666095bdd ]

If tb_drom_read() fails, sw->drom is freed but not set to NULL.  sw->drom
is then freed again in the error path of tb_switch_alloc().

The bug can be triggered by unplugging a thunderbolt device shortly after
it is detected by the thunderbolt driver.

Clear sw->drom if tb_drom_read() fails.

[bhelgaas: add Fixes:, stable versions of interest]
Fixes: 343fcb8c70d7 ("thunderbolt: Fix nontrivial endpoint devices.")
Signed-off-by: Andreas Noever <andreas.noever@gmail.com>
Signed-off-by: Bjorn Helgaas <bhelgaas@google.com>
CC: stable@vger.kernel.org # v3.17+
CC: Lukas Wunner <lukas@wunner.de>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoQE-UART: add "fsl,t1040-ucc-uart" to of_device_id
Zhao Qiang [Wed, 9 Mar 2016 01:48:11 +0000 (09:48 +0800)]
QE-UART: add "fsl,t1040-ucc-uart" to of_device_id

[ Upstream commit 11ca2b7ab432eb90906168c327733575e68d388f ]

New bindings use "fsl,t1040-ucc-uart" as the compatible for qe-uart.
So add it.

Signed-off-by: Zhao Qiang <qiang.zhao@nxp.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoFix OpenSSH pty regression on close
Brian Bloniarz [Sun, 6 Mar 2016 21:16:30 +0000 (13:16 -0800)]
Fix OpenSSH pty regression on close

[ Upstream commit 0f40fbbcc34e093255a2b2d70b6b0fb48c3f39aa ]

OpenSSH expects the (non-blocking) read() of pty master to return
EAGAIN only if it has received all of the slave-side output after
it has received SIGCHLD. This used to work on pre-3.12 kernels.

This fix effectively forces non-blocking read() and poll() to
block for parallel i/o to complete for all ttys. It also unwinds
these changes:

1) f8747d4a466ab2cafe56112c51b3379f9fdb7a12
   tty: Fix pty master read() after slave closes

2) 52bce7f8d4fc633c9a9d0646eef58ba6ae9a3b73
   pty, n_tty: Simplify input processing on final close

3) 1a48632ffed61352a7810ce089dc5a8bcd505a60
   pty: Fix input race when closing

Inspired by analysis and patch from Marc Aurele La France <tsi@tuyoix.net>

Reported-by: Volth <openssh@volth.com>
Reported-by: Marc Aurele La France <tsi@tuyoix.net>
BugLink: https://bugzilla.mindrot.org/show_bug.cgi?id=52
BugLink: https://bugzilla.mindrot.org/show_bug.cgi?id=2492
Signed-off-by: Brian Bloniarz <brian.bloniarz@gmail.com>
Reviewed-by: Peter Hurley <peter@hurleysoftware.com>
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoMIPS: ath79: make bootconsole wait for both THRE and TEMT
Matthias Schiffer [Thu, 24 Mar 2016 15:02:52 +0000 (16:02 +0100)]
MIPS: ath79: make bootconsole wait for both THRE and TEMT

[ Upstream commit f5b556c94c8490d42fea79d7b4ae0ecbc291e69d ]

This makes the ath79 bootconsole behave the same way as the generic 8250
bootconsole.

Also waiting for TEMT (transmit buffer is empty) instead of just THRE
(transmit buffer is not full) ensures that all characters have been
transmitted before the real serial driver starts reconfiguring the serial
controller (which would sometimes result in garbage being transmitted.)
This change does not cause a visible performance loss.

In addition, this seems to fix a hang observed in certain configurations on
many AR7xxx/AR9xxx SoCs during autoconfig of the real serial driver.

A more complete follow-up patch will disable 8250 autoconfig for ath79
altogether (the serial controller is detected as a 16550A, which is not
fully compatible with the ath79 serial, and the autoconfig may lead to
undefined behavior on ath79.)

Cc: <stable@vger.kernel.org>
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoext4: clean up error handling when orphan list is corrupted
Theodore Ts'o [Sat, 30 Apr 2016 04:49:54 +0000 (00:49 -0400)]
ext4: clean up error handling when orphan list is corrupted

[ Upstream commit 7827a7f6ebfcb7f388dc47fddd48567a314701ba ]

Instead of just printing warning messages, if the orphan list is
corrupted, declare the file system is corrupted.  If there are any
reserved inodes in the orphaned inode list, declare the file system
corrupted and stop right away to avoid doing more potential damage to
the file system.

Cc: stable@vger.kernel.org
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoext4: fix hang when processing corrupted orphaned inode list
Theodore Ts'o [Sat, 30 Apr 2016 04:48:54 +0000 (00:48 -0400)]
ext4: fix hang when processing corrupted orphaned inode list

[ Upstream commit c9eb13a9105e2e418f72e46a2b6da3f49e696902 ]

If the orphaned inode list contains inode #5, ext4_iget() returns a
bad inode (since the bootloader inode should never be referenced
directly).  Because of the bad inode, we end up processing the inode
repeatedly and this hangs the machine.

This can be reproduced via:

   mke2fs -t ext4 /tmp/foo.img 100
   debugfs -w -R "ssv last_orphan 5" /tmp/foo.img
   mount -o loop /tmp/foo.img /mnt

(But don't do this if you are using an unpatched kernel if you care
about the system staying functional.  :-)

This bug was found by the port of American Fuzzy Lop into the kernel
to find file system problems[1].  (Since it *only* happens if inode #5
shows up on the orphan list --- 3, 7, 8, etc. won't do it, it's not
surprising that AFL needed two hours before it found it.)

[1] http://events.linuxfoundation.org/sites/events/files/slides/AFL%20filesystem%20fuzzing%2C%20Vault%202016_0.pdf

Cc: stable@vger.kernel.org
Reported by: Vegard Nossum <vegard.nossum@oracle.com>
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoaacraid: Fix for KDUMP driver hang
Raghava Aditya Renukunta [Tue, 26 Apr 2016 06:32:37 +0000 (23:32 -0700)]
aacraid: Fix for KDUMP driver hang

[ Upstream commit 78cbccd3bd683c295a44af8050797dc4a41376ff ]

When KDUMP is triggered the driver first talks to the firmware in INTX
mode, but the adapter firmware is still in MSIX mode. Therefore the first
driver command hangs since the driver is waiting for an INTX response and
firmware gives a MSIX response. If when the OS is installed on a RAID
drive created by the adapter KDUMP will hang since the driver does not
receive a response in sync mode.

Fixed by: Change the firmware to INTX mode if it is in MSIX mode before
sending the first sync command.

Cc: stable@vger.kernel.org
Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoaacraid: Fix for aac_command_thread hang
Raghava Aditya Renukunta [Tue, 26 Apr 2016 06:31:57 +0000 (23:31 -0700)]
aacraid: Fix for aac_command_thread hang

[ Upstream commit fc4bf75ea300a5e62a2419f89dd0e22189dd7ab7 ]

Typically under error conditions, it is possible for aac_command_thread()
to miss the wakeup from kthread_stop() and go back to sleep, causing it
to hang aac_shutdown.

In the observed scenario, the adapter is not functioning correctly and so
aac_fib_send() never completes (or time-outs depending on how it was
called). Shortly after aac_command_thread() starts it performs
aac_fib_send(SendHostTime) which hangs. When aac_probe_one
/aac_get_adapter_info send time outs, kthread_stop is called which breaks
the command thread out of it's hang.

The code will still go back to sleep in schedule_timeout() without
checking kthread_should_stop() so it causes aac_probe_one to hang until
the schedule_timeout() which is 30 minutes.

Fixed by: Adding another kthread_should_stop() before schedule_timeout()
Cc: stable@vger.kernel.org
Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoaacraid: Relinquish CPU during timeout wait
Raghava Aditya Renukunta [Tue, 26 Apr 2016 06:31:26 +0000 (23:31 -0700)]
aacraid: Relinquish CPU during timeout wait

[ Upstream commit 07beca2be24cc710461c0b131832524c9ee08910 ]

aac_fib_send has a special function case for initial commands during
driver initialization using wait < 0(pseudo sync mode). In this case,
the command does not sleep but rather spins checking for timeout.This
loop is calls cpu_relax() in an attempt to allow other processes/threads
to use the CPU, but this function does not relinquish the CPU and so the
command will hog the processor. This was observed in a KDUMP
"crashkernel" and that prevented the "command thread" (which is
responsible for completing the command from being timed out) from
starting because it could not get the CPU.

Fixed by replacing "cpu_relax()" call with "schedule()"
Cc: stable@vger.kernel.org
Signed-off-by: Raghava Aditya Renukunta <RaghavaAditya.Renukunta@microsemi.com>
Reviewed-by: Johannes Thumshirn <jthumshirn@suse.de>
Signed-off-by: Martin K. Petersen <martin.petersen@oracle.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoarm/arm64: KVM: Enforce Break-Before-Make on Stage-2 page tables
Marc Zyngier [Thu, 28 Apr 2016 15:16:31 +0000 (16:16 +0100)]
arm/arm64: KVM: Enforce Break-Before-Make on Stage-2 page tables

[ Upstream commit d4b9e0790aa764c0b01e18d4e8d33e93ba36d51f ]

The ARM architecture mandates that when changing a page table entry
from a valid entry to another valid entry, an invalid entry is first
written, TLB invalidated, and only then the new entry being written.

The current code doesn't respect this, directly writing the new
entry and only then invalidating TLBs. Let's fix it up.

Cc: <stable@vger.kernel.org>
Reported-by: Christoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: Marc Zyngier <marc.zyngier@arm.com>
Signed-off-by: Christoffer Dall <christoffer.dall@linaro.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agostaging: comedi: das1800: fix possible NULL dereference
H Hartley Sweeten [Fri, 8 Apr 2016 17:14:58 +0000 (10:14 -0700)]
staging: comedi: das1800: fix possible NULL dereference

[ Upstream commit d375278d666760e195693b57415ba0a125cadd55 ]

DMA is optional with this driver. If it was not enabled the devpriv->dma
pointer will be NULL.

Fix the possible NULL pointer dereference when trying to disable the DMA
channels in das1800_ai_cancel() and tidy up the comments to fix the
checkpatch.pl issues:
WARNING: line over 80 characters

It's probably harmless in das1800_ai_setup_dma() because the 'desc' pointer
will not be used if DMA is disabled but fix it there also.

Fixes: 99dfc3357e98 ("staging: comedi: das1800: remove depends on ISA_DMA_API limitation")
Cc: <stable@vger.kernel.org> # 4.0+
Signed-off-by: H Hartley Sweeten <hsweeten@visionengravers.com>
Reviewed-by: Ian Abbott <abbotti@mev.co.uk>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoTTY: n_gsm, fix false positive WARN_ON
Jiri Slaby [Tue, 22 Mar 2016 17:09:51 +0000 (18:09 +0100)]
TTY: n_gsm, fix false positive WARN_ON

[ Upstream commit d175feca89a1c162f60f4e3560ca7bc9437c65eb ]

Dmitry reported, that the current cleanup code in n_gsm can trigger a
warning:
WARNING: CPU: 2 PID: 24238 at drivers/tty/n_gsm.c:2048 gsm_cleanup_mux+0x166/0x6b0()
...
Call Trace:
...
 [<ffffffff81247ab9>] warn_slowpath_null+0x29/0x30 kernel/panic.c:490
 [<ffffffff828d0456>] gsm_cleanup_mux+0x166/0x6b0 drivers/tty/n_gsm.c:2048
 [<ffffffff828d4d87>] gsmld_open+0x5b7/0x7a0 drivers/tty/n_gsm.c:2386
 [<ffffffff828b9078>] tty_ldisc_open.isra.2+0x78/0xd0 drivers/tty/tty_ldisc.c:447
 [<ffffffff828b973a>] tty_set_ldisc+0x1ca/0xa70 drivers/tty/tty_ldisc.c:567
 [<     inline     >] tiocsetd drivers/tty/tty_io.c:2650
 [<ffffffff828a14ea>] tty_ioctl+0xb2a/0x2140 drivers/tty/tty_io.c:2883
...

But this is a legal path when open fails to find a space in the
gsm_mux array and tries to clean up. So make it a standard test
instead of a warning.

Reported-by: "Dmitry Vyukov" <dvyukov@google.com>
Cc: Alan Cox <alan@linux.intel.com>
Link: http://lkml.kernel.org/r/CACT4Y+bHQbAB68VFi7Romcs-Z9ZW3kQRvcq+BvHH1oa5NcAdLA@mail.gmail.com
Fixes: 5a640967 ("tty/n_gsm.c: fix a memory leak in gsmld_open()")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoserial: 8250_pci: fix divide error bug if baud rate is 0
David Müller [Wed, 27 Apr 2016 09:58:32 +0000 (11:58 +0200)]
serial: 8250_pci: fix divide error bug if baud rate is 0

[ Upstream commit 6f210c18c1c0f016772c8cd51ae12a02bfb9e7ef ]

Since commit 21947ba654a6 ("serial: 8250_pci: replace switch-case by
formula"), the 8250 driver crashes in the byt_set_termios() function
with a divide error. This is caused by the fact that a baud rate of 0 (B0)
is not handled properly. Fix it by falling back to B9600 in this case.

Signed-off-by: David Müller <d.mueller@elsoft.ch>
Fixes: 21947ba654a6 ("serial: 8250_pci: replace switch-case by formula")
Cc: stable@vger.kernel.org
Suggested-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agousb: core: hub: hub_port_init lock controller instead of bus
Chris Bainbridge [Mon, 25 Apr 2016 12:48:38 +0000 (13:48 +0100)]
usb: core: hub: hub_port_init lock controller instead of bus

[ Upstream commit feb26ac31a2a5cb88d86680d9a94916a6343e9e6 ]

The XHCI controller presents two USB buses to the system - one for USB2
and one for USB3. The hub init code (hub_port_init) is reentrant but
only locks one bus per thread, leading to a race condition failure when
two threads attempt to simultaneously initialise a USB2 and USB3 device:

[    8.034843] xhci_hcd 0000:00:14.0: Timeout while waiting for setup device command
[   13.183701] usb 3-3: device descriptor read/all, error -110

On a test system this failure occurred on 6% of all boots.

The call traces at the point of failure are:

Call Trace:
 [<ffffffff81b9bab7>] schedule+0x37/0x90
 [<ffffffff817da7cd>] usb_kill_urb+0x8d/0xd0
 [<ffffffff8111e5e0>] ? wake_up_atomic_t+0x30/0x30
 [<ffffffff817dafbe>] usb_start_wait_urb+0xbe/0x150
 [<ffffffff817db10c>] usb_control_msg+0xbc/0xf0
 [<ffffffff817d07de>] hub_port_init+0x51e/0xb70
 [<ffffffff817d4697>] hub_event+0x817/0x1570
 [<ffffffff810f3e6f>] process_one_work+0x1ff/0x620
 [<ffffffff810f3dcf>] ? process_one_work+0x15f/0x620
 [<ffffffff810f4684>] worker_thread+0x64/0x4b0
 [<ffffffff810f4620>] ? rescuer_thread+0x390/0x390
 [<ffffffff810fa7f5>] kthread+0x105/0x120
 [<ffffffff810fa6f0>] ? kthread_create_on_node+0x200/0x200
 [<ffffffff81ba183f>] ret_from_fork+0x3f/0x70
 [<ffffffff810fa6f0>] ? kthread_create_on_node+0x200/0x200

Call Trace:
 [<ffffffff817fd36d>] xhci_setup_device+0x53d/0xa40
 [<ffffffff817fd87e>] xhci_address_device+0xe/0x10
 [<ffffffff817d047f>] hub_port_init+0x1bf/0xb70
 [<ffffffff811247ed>] ? trace_hardirqs_on+0xd/0x10
 [<ffffffff817d4697>] hub_event+0x817/0x1570
 [<ffffffff810f3e6f>] process_one_work+0x1ff/0x620
 [<ffffffff810f3dcf>] ? process_one_work+0x15f/0x620
 [<ffffffff810f4684>] worker_thread+0x64/0x4b0
 [<ffffffff810f4620>] ? rescuer_thread+0x390/0x390
 [<ffffffff810fa7f5>] kthread+0x105/0x120
 [<ffffffff810fa6f0>] ? kthread_create_on_node+0x200/0x200
 [<ffffffff81ba183f>] ret_from_fork+0x3f/0x70
 [<ffffffff810fa6f0>] ? kthread_create_on_node+0x200/0x200

Which results from the two call chains:

hub_port_init
 usb_get_device_descriptor
  usb_get_descriptor
   usb_control_msg
    usb_internal_control_msg
     usb_start_wait_urb
      usb_submit_urb / wait_for_completion_timeout / usb_kill_urb

hub_port_init
 hub_set_address
  xhci_address_device
   xhci_setup_device

Mathias Nyman explains the current behaviour violates the XHCI spec:

 hub_port_reset() will end up moving the corresponding xhci device slot
 to default state.

 As hub_port_reset() is called several times in hub_port_init() it
 sounds reasonable that we could end up with two threads having their
 xhci device slots in default state at the same time, which according to
 xhci 4.5.3 specs still is a big no no:

 "Note: Software shall not transition more than one Device Slot to the
  Default State at a time"

 So both threads fail at their next task after this.
 One fails to read the descriptor, and the other fails addressing the
 device.

Fix this in hub_port_init by locking the USB controller (instead of an
individual bus) to prevent simultaneous initialisation of both buses.

Fixes: 638139eb95d2 ("usb: hub: allow to process more usb hub events in parallel")
Link: https://lkml.org/lkml/2016/2/8/312
Link: https://lkml.org/lkml/2016/2/4/748
Signed-off-by: Chris Bainbridge <chris.bainbridge@gmail.com>
Cc: stable <stable@vger.kernel.org>
Acked-by: Mathias Nyman <mathias.nyman@linux.intel.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agobtrfs: bugfix: handle FS_IOC32_{GETFLAGS,SETFLAGS,GETVERSION} in btrfs_ioctl
Luke Dashjr [Thu, 29 Oct 2015 08:22:21 +0000 (08:22 +0000)]
btrfs: bugfix: handle FS_IOC32_{GETFLAGS,SETFLAGS,GETVERSION} in btrfs_ioctl

[ Upstream commit 4c63c2454eff996c5e27991221106eb511f7db38 ]

32-bit ioctl uses these rather than the regular FS_IOC_* versions. They can
be handled in btrfs using the same code. Without this, 32-bit {ch,ls}attr
fail.

Signed-off-by: Luke Dashjr <luke-jr+git@utopios.org>
Cc: stable@vger.kernel.org
Reviewed-by: Josef Bacik <jbacik@fb.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agopinctrl: exynos5440: Use off-stack memory for pinctrl_gpio_range
Andrew Jeffery [Wed, 20 Apr 2016 01:54:17 +0000 (11:24 +0930)]
pinctrl: exynos5440: Use off-stack memory for pinctrl_gpio_range

[ Upstream commit 71324fdc72ef0163e57631aa814a9a81e9e4770b ]

The range is registered into a linked list which can be referenced
throughout the lifetime of the driver. Ensure the range's memory is useful
for the same lifetime by adding it to the driver's private data structure.

The bug was introduced in the driver's initial commit, which was present in
v3.10.

Fixes: f0b9a7e521fa ("pinctrl: exynos5440: add pinctrl driver for Samsung EXYNOS5440 SoC")
Cc: stable@vger.kernel.org
Signed-off-by: Andrew Jeffery <andrew@aj.id.au>
Acked-by: Tomasz Figa <tomasz.figa@gmail.com>
Reviewed-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoath9k: Fix LED polarity for some Mini PCI AR9220 MB92 cards.
Vittorio Gambaletta (VittGam) [Mon, 11 Apr 2016 02:48:55 +0000 (04:48 +0200)]
ath9k: Fix LED polarity for some Mini PCI AR9220 MB92 cards.

[ Upstream commit 0f9edcdd88a993914fa1d1dc369b35dc503979db ]

The Wistron DNMA-92 and Compex WLM200NX have inverted LED polarity
(active high instead of active low).

The same PCI Subsystem ID is used by both cards, which are based on
the same Atheros MB92 design.

Cc: <linux-wireless@vger.kernel.org>
Cc: <ath9k-devel@qca.qualcomm.com>
Cc: <ath9k-devel@lists.ath9k.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Vittorio Gambaletta <linuxbugs@vittgam.net>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoath9k: Add a module parameter to invert LED polarity.
Vittorio Gambaletta (VittGam) [Mon, 11 Apr 2016 02:48:54 +0000 (04:48 +0200)]
ath9k: Add a module parameter to invert LED polarity.

[ Upstream commit cd84042ce9040ad038e958bc67a46fcfc015c736 ]

The LED can be active high instead of active low on some hardware.

Add the led_active_high module parameter. It defaults to -1 to obey
platform data as before.

Setting the parameter to 1 or 0 will force the LED respectively
active high or active low.

Cc: <linux-wireless@vger.kernel.org>
Cc: <ath9k-devel@qca.qualcomm.com>
Cc: <ath9k-devel@lists.ath9k.org>
Cc: <stable@vger.kernel.org>
Signed-off-by: Vittorio Gambaletta <linuxbugs@vittgam.net>
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agocrypto: s5p-sss - Fix missed interrupts when working with 8 kB blocks
Krzysztof Kozlowski [Fri, 22 Apr 2016 12:15:23 +0000 (14:15 +0200)]
crypto: s5p-sss - Fix missed interrupts when working with 8 kB blocks

[ Upstream commit 79152e8d085fd64484afd473ef6830b45518acba ]

The tcrypt testing module on Exynos5422-based Odroid XU3/4 board failed on
testing 8 kB size blocks:

$ sudo modprobe tcrypt sec=1 mode=500
testing speed of async ecb(aes) (ecb-aes-s5p) encryption
test 0 (128 bit key, 16 byte blocks): 21971 operations in 1 seconds (351536 bytes)
test 1 (128 bit key, 64 byte blocks): 21731 operations in 1 seconds (1390784 bytes)
test 2 (128 bit key, 256 byte blocks): 21932 operations in 1 seconds (5614592 bytes)
test 3 (128 bit key, 1024 byte blocks): 21685 operations in 1 seconds (22205440 bytes)
test 4 (128 bit key, 8192 byte blocks):

This was caused by a race issue of missed BRDMA_DONE ("Block cipher
Receiving DMA") interrupt. Device starts processing the data in DMA mode
immediately after setting length of DMA block: receiving (FCBRDMAL) or
transmitting (FCBTDMAL). The driver sets these lengths from interrupt
handler through s5p_set_dma_indata() function (or xxx_setdata()).

However the interrupt handler was first dealing with receive buffer
(dma-unmap old, dma-map new, set receive block length which starts the
operation), then with transmit buffer and finally was clearing pending
interrupts (FCINTPEND). Because of the time window between setting
receive buffer length and clearing pending interrupts, the operation on
receive buffer could end already and driver would miss new interrupt.

User manual for Exynos5422 confirms in example code that setting DMA
block lengths should be the last operation.

The tcrypt hang could be also observed in following blocked-task dmesg:

INFO: task modprobe:258 blocked for more than 120 seconds.
      Not tainted 4.6.0-rc4-next-20160419-00005-g9eac8b7b7753-dirty #42
"echo 0 > /proc/sys/kernel/hung_task_timeout_secs" disables this message.
modprobe        D c06b09d8     0   258    256 0x00000000
[<c06b09d8>] (__schedule) from [<c06b0f24>] (schedule+0x40/0xac)
[<c06b0f24>] (schedule) from [<c06b49f8>] (schedule_timeout+0x124/0x178)
[<c06b49f8>] (schedule_timeout) from [<c06b17fc>] (wait_for_common+0xb8/0x144)
[<c06b17fc>] (wait_for_common) from [<bf0013b8>] (test_acipher_speed+0x49c/0x740 [tcrypt])
[<bf0013b8>] (test_acipher_speed [tcrypt]) from [<bf003e8c>] (do_test+0x2240/0x30ec [tcrypt])
[<bf003e8c>] (do_test [tcrypt]) from [<bf008048>] (tcrypt_mod_init+0x48/0xa4 [tcrypt])
[<bf008048>] (tcrypt_mod_init [tcrypt]) from [<c010177c>] (do_one_initcall+0x3c/0x16c)
[<c010177c>] (do_one_initcall) from [<c0191ff0>] (do_init_module+0x5c/0x1ac)
[<c0191ff0>] (do_init_module) from [<c0185610>] (load_module+0x1a30/0x1d08)
[<c0185610>] (load_module) from [<c0185ab0>] (SyS_finit_module+0x8c/0x98)
[<c0185ab0>] (SyS_finit_module) from [<c01078c0>] (ret_fast_syscall+0x0/0x3c)

Fixes: a49e490c7a8a ("crypto: s5p-sss - add S5PV210 advanced crypto engine support")
Cc: <stable@vger.kernel.org>
Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Tested-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agocrypto: s5p-sss - Remove useless hash interrupt handler
Krzysztof Kozlowski [Tue, 19 Apr 2016 13:44:12 +0000 (15:44 +0200)]
crypto: s5p-sss - Remove useless hash interrupt handler

[ Upstream commit 5512442553bbe8d4fcdba3e17b30f187706384a7 ]

Beside regular feed control interrupt, the driver requires also hash
interrupt for older SoCs (samsung,s5pv210-secss). However after
requesting it, the interrupt handler isn't doing anything with it, not
even clearing the hash interrupt bit.

Driver does not provide hash functions so it is safe to remove the hash
interrupt related code and to not require the interrupt in Device Tree.

Signed-off-by: Krzysztof Kozlowski <k.kozlowski@samsung.com>
Signed-off-by: Herbert Xu <herbert@gondor.apana.org.au>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoPM / Runtime: Fix error path in pm_runtime_force_resume()
Ulf Hansson [Fri, 8 Apr 2016 11:10:23 +0000 (13:10 +0200)]
PM / Runtime: Fix error path in pm_runtime_force_resume()

[ Upstream commit 0ae3aeefabbeef26294e7a349b51f1c761d46c9f ]

As pm_runtime_set_active() may fail because the device's parent isn't
active, we can end up executing the ->runtime_resume() callback for the
device when it isn't allowed.

Fix this by invoking pm_runtime_set_active() before running the callback
and let's also deal with the error code.

Fixes: 37f204164dfb (PM: Add pm_runtime_suspend|resume_force functions)
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Linus Walleij <linus.walleij@linaro.org>
Cc: 3.15+ <stable@vger.kernel.org> # 3.15+
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agopowerpc/book3s64: Fix branching to OOL handlers in relocatable kernel
Hari Bathini [Fri, 15 Apr 2016 12:48:02 +0000 (22:48 +1000)]
powerpc/book3s64: Fix branching to OOL handlers in relocatable kernel

[ Upstream commit 8ed8ab40047a570fdd8043a40c104a57248dd3fd ]

Some of the interrupt vectors on 64-bit POWER server processors are only
32 bytes long (8 instructions), which is not enough for the full
first-level interrupt handler. For these we need to branch to an
out-of-line (OOL) handler. But when we are running a relocatable kernel,
interrupt vectors till __end_interrupts marker are copied down to real
address 0x100. So, branching to labels (ie. OOL handlers) outside this
section must be handled differently (see LOAD_HANDLER()), considering
relocatable kernel, which would need at least 4 instructions.

However, branching from interrupt vector means that we corrupt the
CFAR (come-from address register) on POWER7 and later processors as
mentioned in commit 1707dd16. So, EXCEPTION_PROLOG_0 (6 instructions)
that contains the part up to the point where the CFAR is saved in the
PACA should be part of the short interrupt vectors before we branch out
to OOL handlers.

But as mentioned already, there are interrupt vectors on 64-bit POWER
server processors that are only 32 bytes long (like vectors 0x4f00,
0x4f20, etc.), which cannot accomodate the above two cases at the same
time owing to space constraint. Currently, in these interrupt vectors,
we simply branch out to OOL handlers, without using LOAD_HANDLER(),
which leaves us vulnerable when running a relocatable kernel (eg. kdump
case). While this has been the case for sometime now and kdump is used
widely, we were fortunate not to see any problems so far, for three
reasons:

  1. In almost all cases, production kernel (relocatable) is used for
     kdump as well, which would mean that crashed kernel's OOL handler
     would be at the same place where we end up branching to, from short
     interrupt vector of kdump kernel.
  2. Also, OOL handler was unlikely the reason for crash in almost all
     the kdump scenarios, which meant we had a sane OOL handler from
     crashed kernel that we branched to.
  3. On most 64-bit POWER server processors, page size is large enough
     that marking interrupt vector code as executable (see commit
     429d2e83) leads to marking OOL handler code from crashed kernel,
     that sits right below interrupt vector code from kdump kernel, as
     executable as well.

Let us fix this by moving the __end_interrupts marker down past OOL
handlers to make sure that we also copy OOL handlers to real address
0x100 when running a relocatable kernel.

This fix has been tested successfully in kdump scenario, on an LPAR with
4K page size by using different default/production kernel and kdump
kernel.

Also tested by manually corrupting the OOL handlers in the first kernel
and then kdump'ing, and then causing the OOL handlers to fire - mpe.

Fixes: c1fb6816fb1b ("powerpc: Add relocation on exception vector handlers")
Cc: stable@vger.kernel.org
Signed-off-by: Hari Bathini <hbathini@linux.vnet.ibm.com>
Signed-off-by: Mahesh Salgaonkar <mahesh@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoBluetooth: vhci: Fix race at creating hci device
Takashi Iwai [Thu, 14 Apr 2016 15:32:19 +0000 (17:32 +0200)]
Bluetooth: vhci: Fix race at creating hci device

[ Upstream commit c7c999cb18da88a881e10e07f0724ad0bfaff770 ]

hci_vhci driver creates a hci device object dynamically upon each
HCI_VENDOR_PKT write.  Although it checks the already created object
and returns an error, it's still racy and may build multiple hci_dev
objects concurrently when parallel writes are performed, as the device
tracks only a single hci_dev object.

This patch introduces a mutex to protect against the concurrent device
creations.

Cc: <stable@vger.kernel.org>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agohwmon: (ads7828) Enable internal reference
Akshay Bhat [Mon, 18 Apr 2016 19:47:53 +0000 (15:47 -0400)]
hwmon: (ads7828) Enable internal reference

[ Upstream commit 7a18afe8097731b8ffb6cb5b2b3b418ded77c105 ]

On ads7828 the internal reference defaults to off upon power up. When
using internal reference, it needs to be turned on and the voltage needs
to settle before normal conversion cycle can be started. Hence perform a
dummy read in the probe to enable the internal reference allowing the
voltage to settle before performing a normal read.

Without this fix, the first read from the ADC when using internal
reference always returns incorrect data.

Signed-off-by: Akshay Bhat <akshay.bhat@timesys.com>
Cc: stable@vger.kernel.org # v4.1+
Signed-off-by: Guenter Roeck <linux@roeck-us.net>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agousb: f_mass_storage: test whether thread is running before starting another
Michal Nazarewicz [Fri, 8 Apr 2016 08:24:11 +0000 (10:24 +0200)]
usb: f_mass_storage: test whether thread is running before starting another

[ Upstream commit f78bbcae86e676fad9e6c6bb6cd9d9868ba23696 ]

When binding the function to usb_configuration, check whether the thread
is running before starting another one.  Without that, when function
instance is added to multiple configurations, fsg_bing starts multiple
threads with all but the latest one being forgotten by the driver.  This
leads to obvious thread leaks, possible lockups when trying to halt the
machine and possible more issues.

This fixes issues with legacy/multi¹ gadget as well as configfs gadgets
when mass_storage function is added to multiple configurations.

This change also simplifies API since the legacy gadgets no longer need
to worry about starting the thread by themselves (which was where bug
in legacy/multi was in the first place).

N.B., this patch doesn’t address adding single mass_storage function
instance to a single configuration twice.  Thankfully, there’s no
legitimate reason for such setup plus, if I’m not mistaken, configfs
gadget doesn’t even allow it to be expressed.

¹ I have no example failure though.  Conclusion that legacy/multi has
  a bug is based purely on me reading the code.

Acked-by: Alan Stern <stern@rowland.harvard.edu>
Signed-off-by: Michal Nazarewicz <mina86@mina86.com>
Tested-by: Ivaylo Dimitrov <ivo.g.dimitrov.75@gmail.com>
Cc: Alan Stern <stern@rowland.harvard.edu>
Cc: <stable@vger.kernel.org>
Signed-off-by: Felipe Balbi <felipe.balbi@linux.intel.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agousb: gadget: mass_storage: Use static array for luns
Krzysztof Opasiak [Fri, 31 Jul 2015 11:46:07 +0000 (13:46 +0200)]
usb: gadget: mass_storage: Use static array for luns

[ Upstream commit dd02ea5a33059e4a753ae8bb877b698c54ee2907 ]

This patch replace dynamicly allocated luns array with static one.
This simplifies the code of mass storage function and modules.

Signed-off-by: Krzysztof Opasiak <k.opasiak@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agousb: gadget: mass_storage: Fix freeing luns sysfs implementation
Krzysztof Opasiak [Fri, 31 Jul 2015 11:37:45 +0000 (13:37 +0200)]
usb: gadget: mass_storage: Fix freeing luns sysfs implementation

[ Upstream commit 5542f58c95aef67bc9016855f7c0bd6117b43100 ]

Use device_is_registered() instad of sysfs flag to determine if
we should free sysfs representation of particular LUN.

sysfs flag in fsg common determines if luns attributes should be
exposed using sysfs. This flag is used when creating and freeing
luns. Unfortunately there is no guarantee that this flag will not
be changed between creation and removal of particular LUN. Especially
because of lun.0 which is created during allocating instance of
function. This may lead to resource leak or NULL pointer dereference:

[   62.539925] Unable to handle kernel NULL pointer dereference at virtual address 00000044
[   62.548014] pgd = ec994000
[   62.550679] [00000044] *pgd=6d7be831, *pte=00000000, *ppte=00000000
[   62.556933] Internal error: Oops: 17 [#1] PREEMPT SMP ARM
[   62.562310] Modules linked in: g_mass_storage(+)
[   62.566916] CPU: 2 PID: 613 Comm: insmod Not tainted 4.2.0-rc4-00077-ge29ee91-dirty #125
[   62.574984] Hardware name: SAMSUNG EXYNOS (Flattened Device Tree)
[   62.581061] task: eca56e80 ti: eca76000 task.ti: eca76000
[   62.586450] PC is at kernfs_find_ns+0x8/0xe8
[   62.590698] LR is at kernfs_find_and_get_ns+0x30/0x48
[   62.595732] pc : [<c01277c0>]    lr : [<c0127b88>]    psr: 40010053
[   62.595732] sp : eca77c40  ip : eca77c38  fp : 000008c1
[   62.607187] r10: 00000001  r9 : c0082f38  r8 : ed41ce40
[   62.612395] r7 : c05c1484  r6 : 00000000  r5 : 00000000  r4 : c0814488
[   62.618904] r3 : 00000000  r2 : 00000000  r1 : c05c1484  r0 : 00000000
[   62.625417] Flags: nZcv  IRQs on  FIQs off  Mode SVC_32  ISA ARM  Segment user
[   62.632620] Control: 10c5387d  Table: 6c99404a  DAC: 00000015
[   62.638348] Process insmod (pid: 613, stack limit = 0xeca76210)
[   62.644251] Stack: (0xeca77c40 to 0xeca78000)
[   62.648594] 7c40: c0814488 00000000 00000000 c05c1484 ed41ce40 c0127b88 00000000 c0824888
[   62.656753] 7c60: ed41d038 ed41d030 ed41d000 c012af4c 00000000 c0824858 ed41d038 c02e3314
[   62.664912] 7c80: ed41d030 00000000 ed41ce04 c02d9e8c c070eda8 eca77cb4 000008c1 c058317c
[   62.673071] 7ca0: 000008c1 ed41d030 ed41ce00 ed41ce04 ed41d000 c02da044 ed41cf48 c0375870
[   62.681230] 7cc0: ed9d3c04 ed9d3c00 ed52df80 bf000940 fffffff0 c03758f4 c03758c0 00000000
[   62.689389] 7ce0: bf000564 c03614e0 ed9d3c04 bf000194 c0082f38 00000001 00000000 c0000100
[   62.697548] 7d00: c0814488 c0814488 c086b1dc c05893a8 00000000 ed7e8320 00000000 c0128b88
[   62.705707] 7d20: ed8a6b40 00000000 00000000 ed410500 ed8a6b40 c0594818 ed7e8320 00000000
[   62.713867] 7d40: 00000000 c0129f20 00000000 c082c444 ed8a6b40 c012a684 00001000 00000000
[   62.722026] 7d60: c0594818 c082c444 00000000 00000000 ed52df80 ed52df80 00000000 00000000
[   62.730185] 7d80: 00000000 00000000 00000001 00000002 ed8e9b70 ed52df80 bf0006d0 00000000
[   62.738345] 7da0: ed8e9b70 ed410500 ed618340 c036129c ed8c1c00 bf0006d0 c080b158 ed8c1c00
[   62.746504] 7dc0: bf0006d0 c080b158 ed8c1c08 ed410500 c0082f38 ed618340 000008c1 c03640ac
[   62.754663] 7de0: 00000000 bf0006d0 c082c8dc c080b158 c080b158 c03642d4 00000000 bf003000
[   62.762822] 7e00: 00000000 c0009784 00000000 00000001 00000000 c05849b0 00000002 ee7ab780
[   62.770981] 7e20: 00000002 ed4105c0 0000c53e 000000d0 c0808600 eca77e5c 00000004 00000000
[   62.779140] 7e40: bf000000 c0095680 c08075a0 ee001f00 ed4105c0 c00cadc0 ed52df80 bf000780
[   62.787300] 7e60: ed4105c0 bf000780 00000001 bf0007c8 c0082f38 ed618340 000008c1 c0083e24
[   62.795459] 7e80: 00000001 bf000780 00000001 eca77f58 00000001 bf000780 00000001 c00857f4
[   62.803618] 7ea0: bf00078c 00007fff 00000000 c00835b4 eca77f58 00000000 c0082fac eca77f58
[   62.811777] 7ec0: f05038c0 0003b008 bf000904 00000000 00000000 bf00078c 6e72656b 00006c65
[   62.819936] 7ee0: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[   62.828095] 7f00: 00000000 00000000 00000000 00000000 00000000 00000000 00000000 00000000
[   62.836255] 7f20: 00000000 00000000 00000000 00000000 00000000 00000000 00000003 0003b008
[   62.844414] 7f40: 0000017b c000f5c8 eca76000 00000000 0003b008 c0085df8 f04ef000 0001b8a9
[   62.852573] 7f60: f0503258 f05030c2 f0509fe8 00000968 00000dc8 00000000 00000000 00000000
[   62.860732] 7f80: 00000029 0000002a 00000011 00000000 0000000a 00000000 33f6eb00 0003b008
[   62.868892] 7fa0: bef01cac c000f400 33f6eb00 0003b008 00000003 0003b008 00000000 00000003
[   62.877051] 7fc0: 33f6eb00 0003b008 bef01cac 0000017b 00000000 0003b008 0000000b 0003b008
[   62.885210] 7fe0: bef01ae0 bef01ad0 0001dc23 b6e8c162 800b0070 00000003 c0c0c0c0 c0c0c0c0
[   62.893380] [<c01277c0>] (kernfs_find_ns) from [<c0824888>] (pm_qos_latency_tolerance_attr_group+0x0/0x10)
[   62.903005] Code: e28dd00c e8bd80f0 e92d41f0 e2923000 (e1d0e4b4)
[   62.909115] ---[ end trace 02fb4373ef095c7b ]---

Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Krzysztof Opasiak <k.opasiak@samsung.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agousb: gadget: mass_storage: Free buffers if create lun fails
Krzysztof Opasiak [Mon, 20 Jul 2015 18:15:17 +0000 (20:15 +0200)]
usb: gadget: mass_storage: Free buffers if create lun fails

[ Upstream commit 903588a99c26ed6f103eaa0cfa4ccbe9cd779398 ]

Creation of LUN 0 may fail (for example due to ENOMEM).
As fsg_common_set_num_buffers() does some memory allocation
we should free it before it becomes unavailable.

Signed-off-by: Krzysztof Opasiak <k.opasiak@samsung.com>
Acked-by: Michal Nazarewicz <mina86@mina86.com>
Signed-off-by: Felipe Balbi <balbi@ti.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years ago[media] media: v4l2-compat-ioctl32: fix missing reserved field copy in put_v4l2_create32
Tiffany Lin [Mon, 14 Mar 2016 11:16:14 +0000 (08:16 -0300)]
[media] media: v4l2-compat-ioctl32: fix missing reserved field copy in put_v4l2_create32

[ Upstream commit baf43c6eace43868e490f18560287fa3481b2159 ]

In v4l2-compliance utility, test VIDIOC_CREATE_BUFS will check whether reserved
filed of v4l2_create_buffers filled with zero
Reserved field is filled with zero in v4l_create_bufs.
This patch copy reserved field of v4l2_create_buffer from kernel space to user
space

Signed-off-by: Tiffany Lin <tiffany.lin@mediatek.com>
Signed-off-by: Hans Verkuil <hans.verkuil@cisco.com>
Cc: <stable@vger.kernel.org> # for v3.19 and up
Signed-off-by: Mauro Carvalho Chehab <mchehab@osg.samsung.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agomfd: intel_quark_i2c_gpio: load gpio driver first
Andy Shevchenko [Fri, 23 Oct 2015 09:16:42 +0000 (12:16 +0300)]
mfd: intel_quark_i2c_gpio: load gpio driver first

[ Upstream commit c39dc960e9c5196022cbc46507d4782b6fc314f6 ]

On Intel Galileo boards the GPIO expander is connected to i2c bus. Moreover it
is able to generate interrupt, but interrupt line is connected to GPIO. That's
why we have to have GPIO driver in place when we will probe i2c host with
device connected to it.

Acked-by: Lee Jones <lee.jones@linaro.org>
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Signed-off-by: Wolfram Sang <wsa@the-dreams.de>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agocpuidle: Indicate when a device has been unregistered
Dave Gerlach [Tue, 5 Apr 2016 19:05:38 +0000 (14:05 -0500)]
cpuidle: Indicate when a device has been unregistered

[ Upstream commit c998c07836f985b24361629dc98506ec7893e7a0 ]

Currently the 'registered' member of the cpuidle_device struct is set
to 1 during cpuidle_register_device. In this same function there are
checks to see if the device is already registered to prevent duplicate
calls to register the device, but this value is never set to 0 even on
unregister of the device. Because of this, any attempt to call
cpuidle_register_device after a call to cpuidle_unregister_device will
fail which shouldn't be the case.

To prevent this, set registered to 0 when the device is unregistered.

Fixes: c878a52d3c7c (cpuidle: Check if device is already registered)
Signed-off-by: Dave Gerlach <d-gerlach@ti.com>
Acked-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: All applicable <stable@vger.kernel.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoBluetooth: vhci: purge unhandled skbs
Jiri Slaby [Sat, 19 Mar 2016 10:49:43 +0000 (11:49 +0100)]
Bluetooth: vhci: purge unhandled skbs

[ Upstream commit 13407376b255325fa817798800117a839f3aa055 ]

The write handler allocates skbs and queues them into data->readq.
Read side should read them, if there is any. If there is none, skbs
should be dropped by hdev->flush. But this happens only if the device
is HCI_UP, i.e. hdev->power_on work was triggered already. When it was
not, skbs stay allocated in the queue when /dev/vhci is closed. So
purge the queue in ->release.

Program to reproduce:
#include <err.h>
#include <fcntl.h>
#include <stdio.h>
#include <unistd.h>

#include <sys/stat.h>
#include <sys/types.h>
#include <sys/uio.h>

int main()
{
char buf[] = { 0xff, 0 };
struct iovec iov = {
.iov_base = buf,
.iov_len = sizeof(buf),
};
int fd;

while (1) {
fd = open("/dev/vhci", O_RDWR);
if (fd < 0)
err(1, "open");

usleep(50);

if (writev(fd, &iov, 1) < 0)
err(1, "writev");

usleep(50);

close(fd);
}

return 0;
}

Result:
kmemleak: 4609 new suspected memory leaks
unreferenced object 0xffff88059f4d5440 (size 232):
  comm "vhci", pid 1084, jiffies 4294912542 (age 37569.296s)
  hex dump (first 32 bytes):
    20 f0 23 87 05 88 ff ff 20 f0 23 87 05 88 ff ff   .#..... .#.....
    00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
  backtrace:
...
    [<ffffffff81ece010>] __alloc_skb+0x0/0x5a0
    [<ffffffffa021886c>] vhci_create_device+0x5c/0x580 [hci_vhci]
    [<ffffffffa0219436>] vhci_write+0x306/0x4c8 [hci_vhci]

Fixes: 23424c0d31 (Bluetooth: Add support creating virtual AMP controllers)
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Cc: stable 3.13+ <stable@vger.kernel.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoBluetooth: vhci: fix open_timeout vs. hdev race
Jiri Slaby [Sat, 19 Mar 2016 10:05:18 +0000 (11:05 +0100)]
Bluetooth: vhci: fix open_timeout vs. hdev race

[ Upstream commit 373a32c848ae3a1c03618517cce85f9211a6facf ]

Both vhci_get_user and vhci_release race with open_timeout work. They
both contain cancel_delayed_work_sync, but do not test whether the
work actually created hdev or not. Since the work can be in progress
and _sync will wait for finishing it, we can have data->hdev allocated
when cancel_delayed_work_sync returns. But the call sites do 'if
(data->hdev)' *before* cancel_delayed_work_sync.

As a result:
* vhci_get_user allocates a second hdev and puts it into
  data->hdev. The former is leaked.
* vhci_release does not release data->hdev properly as it thinks there
  is none.

Fix both cases by moving the actual test *after* the call to
cancel_delayed_work_sync.

This can be hit by this program:
#include <err.h>
#include <fcntl.h>
#include <stdio.h>
#include <stdlib.h>
#include <time.h>
#include <unistd.h>

#include <sys/stat.h>
#include <sys/types.h>

int main(int argc, char **argv)
{
int fd;

srand(time(NULL));

while (1) {
const int delta = (rand() % 200 - 100) * 100;

fd = open("/dev/vhci", O_RDWR);
if (fd < 0)
err(1, "open");

usleep(1000000 + delta);

close(fd);
}

return 0;
}

And the result is:
BUG: KASAN: use-after-free in skb_queue_tail+0x13e/0x150 at addr ffff88006b0c1228
Read of size 8 by task kworker/u13:1/32068
=============================================================================
BUG kmalloc-192 (Tainted: G            E     ): kasan: bad access detected
-----------------------------------------------------------------------------

Disabling lock debugging due to kernel taint
INFO: Allocated in vhci_open+0x50/0x330 [hci_vhci] age=260 cpu=3 pid=32040
...
kmem_cache_alloc_trace+0x150/0x190
vhci_open+0x50/0x330 [hci_vhci]
misc_open+0x35b/0x4e0
chrdev_open+0x23b/0x510
...
INFO: Freed in vhci_release+0xa4/0xd0 [hci_vhci] age=9 cpu=2 pid=32040
...
__slab_free+0x204/0x310
vhci_release+0xa4/0xd0 [hci_vhci]
...
INFO: Slab 0xffffea0001ac3000 objects=16 used=13 fp=0xffff88006b0c1e00 flags=0x5fffff80004080
INFO: Object 0xffff88006b0c1200 @offset=4608 fp=0xffff88006b0c0600
Bytes b4 ffff88006b0c11f0: 09 df 00 00 01 00 00 00 00 00 00 00 00 00 00 00  ................
Object ffff88006b0c1200: 00 06 0c 6b 00 88 ff ff 00 00 00 00 00 00 00 00  ...k............
Object ffff88006b0c1210: 10 12 0c 6b 00 88 ff ff 10 12 0c 6b 00 88 ff ff  ...k.......k....
Object ffff88006b0c1220: c0 46 c2 6b 00 88 ff ff c0 46 c2 6b 00 88 ff ff  .F.k.....F.k....
Object ffff88006b0c1230: 01 00 00 00 01 00 00 00 e0 ff ff ff 0f 00 00 00  ................
Object ffff88006b0c1240: 40 12 0c 6b 00 88 ff ff 40 12 0c 6b 00 88 ff ff  @..k....@..k....
Object ffff88006b0c1250: 50 0d 6e a0 ff ff ff ff 00 02 00 00 00 00 ad de  P.n.............
Object ffff88006b0c1260: 00 00 00 00 00 00 00 00 ab 62 02 00 01 00 00 00  .........b......
Object ffff88006b0c1270: 90 b9 19 81 ff ff ff ff 38 12 0c 6b 00 88 ff ff  ........8..k....
Object ffff88006b0c1280: 03 00 20 00 ff ff ff ff ff ff ff ff 00 00 00 00  .. .............
Object ffff88006b0c1290: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
Object ffff88006b0c12a0: 00 00 00 00 00 00 00 00 00 80 cd 3d 00 88 ff ff  ...........=....
Object ffff88006b0c12b0: 00 20 00 00 00 00 00 00 00 00 00 00 00 00 00 00  . ..............
Redzone ffff88006b0c12c0: bb bb bb bb bb bb bb bb                          ........
Padding ffff88006b0c13f8: 00 00 00 00 00 00 00 00                          ........
CPU: 3 PID: 32068 Comm: kworker/u13:1 Tainted: G    B       E      4.4.6-0-default #1
Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS rel-1.8.1-0-g4adadbd-20151112_172657-sheep25 04/01/2014
Workqueue: hci0 hci_cmd_work [bluetooth]
 00000000ffffffff ffffffff81926cfa ffff88006be37c68 ffff88006bc27180
 ffff88006b0c1200 ffff88006b0c1234 ffffffff81577993 ffffffff82489320
 ffff88006bc24240 0000000000000046 ffff88006a100000 000000026e51eb80
Call Trace:
...
 [<ffffffff81ec8ebe>] ? skb_queue_tail+0x13e/0x150
 [<ffffffffa06e027c>] ? vhci_send_frame+0xac/0x100 [hci_vhci]
 [<ffffffffa0c61268>] ? hci_send_frame+0x188/0x320 [bluetooth]
 [<ffffffffa0c61515>] ? hci_cmd_work+0x115/0x310 [bluetooth]
 [<ffffffff811a1375>] ? process_one_work+0x815/0x1340
 [<ffffffff811a1f85>] ? worker_thread+0xe5/0x11f0
 [<ffffffff811a1ea0>] ? process_one_work+0x1340/0x1340
 [<ffffffff811b3c68>] ? kthread+0x1c8/0x230
...
Memory state around the buggy address:
 ffff88006b0c1100: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
 ffff88006b0c1180: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc
>ffff88006b0c1200: fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb fb
                                  ^
 ffff88006b0c1280: fb fb fb fb fb fb fb fb fc fc fc fc fc fc fc fc
 ffff88006b0c1300: fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc fc

Fixes: 23424c0d31 (Bluetooth: Add support creating virtual AMP controllers)
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: stable 3.13+ <stable@vger.kernel.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agodrm/gma500: Fix possible out of bounds read
Itai Handler [Mon, 2 Nov 2015 22:20:56 +0000 (00:20 +0200)]
drm/gma500: Fix possible out of bounds read

[ Upstream commit 7ccca1d5bf69fdd1d3c5fcf84faf1659a6e0ad11 ]

Fix possible out of bounds read, by adding missing comma.
The code may read pass the end of the dsi_errors array
when the most significant bit (bit #31) in the intr_stat register
is set.
This bug has been detected using CppCheck (static analysis tool).

Cc: stable@vger.kernel.org
Signed-off-by: Itai Handler <itai_handler@hotmail.com>
Signed-off-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agortlwifi: btcoexist: Implement antenna selection
Larry Finger [Wed, 16 Mar 2016 18:33:35 +0000 (13:33 -0500)]
rtlwifi: btcoexist: Implement antenna selection

[ Upstream commit baa1702290953295e421f0f433e2b1ff4815827c ]

The previous patch added an option to rtl8723be to manually select the
antenna for those cases when only a single antenna is present, and the
on-board EEPROM is incorrectly programmed. This patch implements the
necessary changes in the Bluetooth coexistence driver.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org> [V4.0+]
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agortlwifi: rtl8723be: Add antenna select module parameter
Larry Finger [Wed, 16 Mar 2016 18:33:34 +0000 (13:33 -0500)]
rtlwifi: rtl8723be: Add antenna select module parameter

[ Upstream commit c18d8f5095715c56bb3cd9cba64242542632054b ]

A number of new laptops have been delivered with only a single antenna.
In principle, this is OK; however, a problem arises when the on-board
EEPROM is programmed to use the other antenna connection. The option
of opening the computer and moving the connector is not always possible
as it will void the warranty in some cases. In addition, this solution
breaks the Windows driver when the box dual boots Linux and Windows.

A fix involving a new module parameter has been developed.  This commit
adds the new parameter and implements the changes needed for the driver.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org> [V4.0+]
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agortlwifi: rtl8723be: Fix module parameter initialization
Sasha Levin [Mon, 30 May 2016 22:24:55 +0000 (18:24 -0400)]
rtlwifi: rtl8723be: Fix module parameter initialization

[ Upstream commit 7079604ddb83f428359feace3aeaf8a9f435be4a ]

This driver has a number of errors in the module initialization. These
include the following:

Parameter msi_support is stored in two places - one is removed.
Paramters sw_crypto and disable_watchdog were never stored in the final
locations, nor were they initialized properly.

Signed-off-by: Larry Finger <Larry.Finger@lwfinger.net>
Cc: Stable <stable@vger.kernel.org>
Signed-off-by: Kalle Valo <kvalo@codeaurora.org>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agolibxfs: pack the agfl header structure so XFS_AGFL_SIZE is correct
Darrick J. Wong [Mon, 4 Jan 2016 05:13:21 +0000 (16:13 +1100)]
libxfs: pack the agfl header structure so XFS_AGFL_SIZE is correct

[ Upstream commit 96f859d52bcb1c6ea6f3388d39862bf7143e2f30 ]

Because struct xfs_agfl is 36 bytes long and has a 64-bit integer
inside it, gcc will quietly round the structure size up to the nearest
64 bits -- in this case, 40 bytes.  This results in the XFS_AGFL_SIZE
macro returning incorrect results for v5 filesystems on 64-bit
machines (118 items instead of 119).  As a result, a 32-bit xfs_repair
will see garbage in AGFL item 119 and complain.

Therefore, tell gcc not to pad the structure so that the AGFL size
calculation is correct.

cc: <stable@vger.kernel.org> # 3.10 - 4.4
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoxfs: Don't wrap growfs AGFL indexes
Dave Chinner [Tue, 5 Apr 2016 21:06:20 +0000 (07:06 +1000)]
xfs: Don't wrap growfs AGFL indexes

[ Upstream commit ad747e3b299671e1a53db74963cc6c5f6cdb9f6d ]

Commit 96f859d ("libxfs: pack the agfl header structure so
XFS_AGFL_SIZE is correct") allowed the freelist to use the empty
slot at the end of the freelist on 64 bit systems that was not
being used due to sizeof() rounding up the structure size.

This has caused versions of xfs_repair prior to 4.5.0 (which also
has the fix) to report this as a corruption once the filesystem has
been grown. Older kernels can also have problems (seen from a whacky
container/vm management environment) mounting filesystems grown on a
system with a newer kernel than the vm/container it is deployed on.

To avoid this problem, change the initial free list indexes not to
wrap across the end of the AGFL, hence avoiding the initialisation
of agf_fllast to the last index in the AGFL.

cc: <stable@vger.kernel.org> # 4.4-4.5
Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Carlos Maiolino <cmaiolino@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoxfs: disallow rw remount on fs with unknown ro-compat features
Eric Sandeen [Tue, 5 Apr 2016 21:05:41 +0000 (07:05 +1000)]
xfs: disallow rw remount on fs with unknown ro-compat features

[ Upstream commit d0a58e833931234c44e515b5b8bede32bd4e6eed ]

Today, a kernel which refuses to mount a filesystem read-write
due to unknown ro-compat features can still transition to read-write
via the remount path.  The old kernel is most likely none the wiser,
because it's unaware of the new feature, and isn't using it.  However,
writing to the filesystem may well corrupt metadata related to that
new feature, and moving to a newer kernel which understand the feature
will have problems.

Right now the only ro-compat feature we have is the free inode btree,
which showed up in v3.16.  It would be good to push this back to
all the active stable kernels, I think, so that if anyone is using
newer mkfs (which enables the finobt feature) with older kernel
releases, they'll be protected.

Cc: <stable@vger.kernel.org> # 3.10.x-
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Bill O'Donnell <billodo@redhat.com>
Reviewed-by: Dave Chinner <dchinner@redhat.com>
Signed-off-by: Dave Chinner <david@fromorbit.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoath5k: Change led pin configuration for compaq c700 laptop
Joseph Salisbury [Mon, 14 Mar 2016 18:51:48 +0000 (14:51 -0400)]
ath5k: Change led pin configuration for compaq c700 laptop

[ Upstream commit 7b9bc799a445aea95f64f15e0083cb19b5789abe ]

BugLink: http://bugs.launchpad.net/bugs/972604
Commit 09c9bae26b0d3c9472cb6ae45010460a2cee8b8d ("ath5k: add led pin
configuration for compaq c700 laptop") added a pin configuration for the Compaq
c700 laptop.  However, the polarity of the led pin is reversed.  It should be
red for wifi off and blue for wifi on, but it is the opposite.  This bug was
reported in the following bug report:
http://pad.lv/972604

Fixes: 09c9bae26b0d3c9472cb6ae45010460a2cee8b8d ("ath5k: add led pin configuration for compaq c700 laptop")
Signed-off-by: Joseph Salisbury <joseph.salisbury@canonical.com>
Cc: stable@vger.kernel.org
Signed-off-by: Kalle Valo <kvalo@qca.qualcomm.com>
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
8 years agoregulator: Try to resolve regulators supplies on registration
Javier Martinez Canillas [Wed, 23 Mar 2016 23:59:34 +0000 (20:59 -0300)]
regulator: Try to resolve regulators supplies on registration

[ Upstream commit 5e3ca2b349b1e2c80b060b51bbf2af37448fad85 ]

Commit 6261b06de565 ("regulator: Defer lookup of supply to regulator_get")
moved the regulator supplies lookup logic from the regulators registration
to the regulators get time.

Unfortunately, that changed the behavior of the regulator core since now a
parent supply with a child regulator marked as always-on, won't be enabled
unless a client driver attempts to get the child regulator during boot.

This patch tries to resolve the parent supply for the already registered
regulators each time that a new regulator is registered. So the regulators
that have child regulators marked as always on will be enabled regardless
if a driver gets the child regulator or not.

That was the behavior before the mentioned commit, since parent supplies
were looked up at regulator registration time instead of during child get.

Since regulator_resolve_supply() checks for rdev->supply, most of the times
it will be a no-op. Errors aren't checked to keep the possible out of order
dependencies which was the motivation for the mentioned commit.

Also, the supply being available will be enforced on regulator get anyways
in case the resolve fails on regulators registration.

Fixes: 6261b06de565 ("regulator: Defer lookup of supply to regulator_get")
Suggested-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
Signed-off-by: Mark Brown <broonie@kernel.org>
Cc: <stable@vger.kernel.org> # 4.1+
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>