Commit graph

72594 commits

Author SHA1 Message Date
Dan Carpenter
2d29748003 x86, microcode, AMD: Fix use after free in free_cache()
list_for_each_entry_reverse() dereferences the iterator, but we already
freed it. I don't see a reason that this has to be done in reverse order
so change it to use list_for_each_entry_safe().

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
2012-09-19 18:06:25 +02:00
Peter Senna Tschudin
4b8073e467 arch/x86: Remove unecessary semicolons
Found by http://coccinelle.lip6.fr/

Signed-off-by: Peter Senna Tschudin <peter.senna@gmail.com>
Cc: avi@redhat.com
Cc: mtosatti@redhat.com
Cc: a.p.zijlstra@chello.nl
Cc: rusty@rustcorp.com.au
Cc: masami.hiramatsu.pt@hitachi.com
Cc: suresh.b.siddha@intel.com
Cc: joerg.roedel@amd.com
Cc: agordeev@redhat.com
Cc: yinghai@kernel.org
Cc: bhelgaas@google.com
Cc: liuj97@gmail.com
Link: http://lkml.kernel.org/r/1347986174-30287-7-git-send-email-peter.senna@gmail.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-09-19 17:32:48 +02:00
Ingo Molnar
81a15f2ee5 Linux 3.6-rc6
-----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2.0.18 (GNU/Linux)
 
 iQEcBAABAgAGBQJQVkutAAoJEHm+PkMAQRiGW8sH/36FVQ3zI75QH16AmR++2nMZ
 BRJGoxcRFMssrXTYVdkMyzygf8b7MZbNEn1qt2g63MNzGaJucPlw5NVL4GLzR+zr
 x/EglLrTEPCD5el9wJ3ls9iC1soudKQTvC2BjcdUjpoSwHrDM/7GKfbOacE54Wqc
 C1VHCcg5DWOD7F0RnYT2SQEVCeDODNmcyFdk7Oi4cUicTPJoYWJ9O9MGfBDBok0N
 M+dXxa9nvsl7EeEKpBKH9vo4TfXn3Gsj6LCRdedvI15ilZjfo8jdHYbSn7KBfQuZ
 JIKRnqkaQ1JfMFt+M/JJZ1b/+Wrd4HLMmmn5oUmrGGIvhpi32nJfi/97+nSy8iU=
 =c5gW
 -----END PGP SIGNATURE-----

Merge tag 'v3.6-rc6' into x86/cleanups

Merge Linux v3.6-rc6 before applying more cleanups.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-09-19 17:32:12 +02:00
Stephane Eranian
20a36e39d5 perf/x86: Fix Intel Ivy Bridge support
This patch updates the existing Intel IvyBridge (model 58)
support with proper PEBS event constraints. It cannot reuse
the same as SandyBridge because some events (0xd3) are
specific to IvyBridge.

Also there is no UOPS_DISPATCHED.THREAD on IVB, so do not
populate the PERF_COUNT_HW_STALLED_CYCLES_BACKEND mapping.

Signed-off-by: Stephane Eranian <eranian@google.com>
Cc: peterz@infradead.org
Cc: ak@linux.intel.com
Link: http://lkml.kernel.org/r/20120910230701.GA5898@quad
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-09-19 17:28:47 +02:00
Jan Beulich
e26a44a2d6 x86: Use REP BSF unconditionally
Make "REP BSF" unconditional, as per the suggestion of hpa
and Linus, this removes the insane BSF_PREFIX conditional
and simplifies the logic.

Suggested-by: "H. Peter Anvin" <hpa@zytor.com>
Suggested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Jan Beulich <jbeulich@suse.com>
Link: http://lkml.kernel.org/r/5058741E020000780009C014@nat28.tlf.novell.com
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-09-19 17:26:08 +02:00
Borislav Petkov
924e101a7a x86/debug: Dump family, model, stepping of the boot CPU
When acting on a user bug report, we find ourselves constantly
asking for /proc/cpuinfo in order to know the exact family,
model, stepping of the CPU in question.

Instead of having to ask this, add this to dmesg so that it is
visible and no ambiguities can ensue from looking at the
official name string of the CPU coming from CPUID and trying
to map it to f/m/s.

Output then looks like this:

[    0.146041] smpboot: CPU0: AMD FX(tm)-8100 Eight-Core Processor (fam: 15, model: 01, stepping: 02)

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
Cc: Andreas Herrmann <andreas.herrmann3@amd.com>
Link: http://lkml.kernel.org/r/1347640666-13638-1-git-send-email-bp@amd64.org
[ tweaked it minimally to add commas. ]
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-09-19 17:12:01 +02:00
Dan Carpenter
1e6dd8adc7 perf: Fix off by one test in perf_reg_value()
The test should be >= ARRAY_SIZE() instead of > ARRAY_SIZE().

Signed-off-by: Dan Carpenter <dan.carpenter@oracle.com>
Acked-by: Jiri Olsa <jolsa@redhat.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Frederic Weisbecker <fweisbec@gmail.com>
Cc: Arnaldo Carvalho de Melo <acme@redhat.com>
Link: http://lkml.kernel.org/r/20120905123126.GC6128@elgon.mountain
Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-09-19 17:08:40 +02:00
Ingo Molnar
d0616c1775 Merge branch 'uprobes/core' of git://git.kernel.org/pub/scm/linux/kernel/git/oleg/misc into perf/core
Pull uprobes fixes + cleanups from Oleg Nesterov.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-09-19 17:03:07 +02:00
Ingo Molnar
961ebea4ae Patch 1/2 which enables MCA by default because I still see bugreports
where CONFIG_X86_MCE is disabled and this is a bad idea so turning it on
 by default makes sense to me. The second one is a trivial cleanup.
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1.4.12 (GNU/Linux)
 
 iQIcBAABAgAGBQJQV2BDAAoJEBLB8Bhh3lVKbfcP/AkfNP5rbK5P5bkoD8Y4jc0r
 XFfTd//96sjN9SvcZyO/dVnO3S8600AJXfjUs+Amd9horaUOdzf1X8xfOKmvupbL
 MqG9WKOaIdgCdXfXvh6S2vNySNErybLWLaxpm8n/7xjQhumVzo1TU41mMzcU5Ctt
 OmhGD2hjF7DXSVDIirDs8XBGyr3YbVRN1Dj/FQDZ1rjOEjAMNsQnZVthFdhHOmfN
 YKaGdP+WbI38Jm3RGEDi3mcnv63A4jL6G4IRDEnsFg45ull1HjTN1tF6lz8tbwDc
 M36jZRM3sqMYDk5P6a9qtnuNTe51Sl9HkbA9Ym0+5hrWua/c16Z6Bt5+NgoDO61O
 LlCU0FZNtmJujqC2oUdVWovpr0z1gnZwlbcvVIPEKVnln5iZC5PkPSxZg7Xn6K7e
 1mMpWA6y9vsF32qxZuD2XAF32/muVgz0zFbVQRgKH6YS064PB2HWgr9ZU0XtVLW9
 9Tfrafqg0CQO/o74oAnbjjg87O4C5p53elpSa/4+4Hwn9E2M+6lKjVyAWNzKv2r3
 +POT4O1IgOE7f2WnPsR1w1F3P00e65xE3IC0JnZyRYh1P8P1FEnVWklGxi/y+dNN
 qVzNDXE5ZL+fKEs2WbP9gr6lTDB0riGc93p9iycDQKVpVq5mePe34CZk/SEZzoGV
 CH/lZMXMrkYqef5wNaAT
 =ywsW
 -----END PGP SIGNATURE-----

Merge tag 'ras_queue_for_3.7' of git://git.kernel.org/pub/scm/linux/kernel/git/ras/ras into x86/mce

Pull MCE changes from Borislav Petkov:

 " Patch 1/2 which enables MCA by default because I still see bugreports
   where CONFIG_X86_MCE is disabled and this is a bad idea so turning it on
   by default makes sense to me. The second one is a trivial cleanup. "

Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-09-19 17:01:50 +02:00
Ingo Molnar
f1f6524476 Linux 3.6-rc6
-----BEGIN PGP SIGNATURE-----
 Version: GnuPG v2.0.18 (GNU/Linux)
 
 iQEcBAABAgAGBQJQVkutAAoJEHm+PkMAQRiGW8sH/36FVQ3zI75QH16AmR++2nMZ
 BRJGoxcRFMssrXTYVdkMyzygf8b7MZbNEn1qt2g63MNzGaJucPlw5NVL4GLzR+zr
 x/EglLrTEPCD5el9wJ3ls9iC1soudKQTvC2BjcdUjpoSwHrDM/7GKfbOacE54Wqc
 C1VHCcg5DWOD7F0RnYT2SQEVCeDODNmcyFdk7Oi4cUicTPJoYWJ9O9MGfBDBok0N
 M+dXxa9nvsl7EeEKpBKH9vo4TfXn3Gsj6LCRdedvI15ilZjfo8jdHYbSn7KBfQuZ
 JIKRnqkaQ1JfMFt+M/JJZ1b/+Wrd4HLMmmn5oUmrGGIvhpi32nJfi/97+nSy8iU=
 =c5gW
 -----END PGP SIGNATURE-----

Merge tag 'v3.6-rc6' into x86/mce

Merge Linux v3.6-rc6, to refresh this tree.

Signed-off-by: Ingo Molnar <mingo@kernel.org>
2012-09-19 17:01:25 +02:00
Suresh Siddha
a8615af4bc x86, fpu: remove cpu_has_xmm check in the fx_finit()
CPUs with FXSAVE but no XMM/MXCSR (Pentium II from Intel,
Crusoe/TM-3xxx/5xxx from Transmeta, and presumably some of the K6
generation from AMD) ever looked at the mxcsr field during
fxrstor/fxsave. So remove the cpu_has_xmm check in the fx_finit()

Reported-by: Al Viro <viro@zeniv.linux.org.uk>
Acked-by: H. Peter Anvin <hpa@zytor.com>
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1347300665-6209-6-git-send-email-suresh.b.siddha@intel.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:52:24 -07:00
Suresh Siddha
e00229819f x86, fpu: make eagerfpu= boot param tri-state
Add the "eagerfpu=auto" (that selects the default scheme in
enabling eagerfpu) which can override compiled-in boot parameters
like "eagerfpu=on/off" (that force enable/disable eagerfpu).

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1347300665-6209-5-git-send-email-suresh.b.siddha@intel.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:52:24 -07:00
Suresh Siddha
212b02125f x86, fpu: enable eagerfpu by default for xsaveopt
xsaveopt/xrstor support optimized state save/restore by tracking the
INIT state and MODIFIED state during context-switch.

Enable eagerfpu by default for processors supporting xsaveopt.
Can be disabled by passing "eagerfpu=off" boot parameter.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1347300665-6209-3-git-send-email-suresh.b.siddha@intel.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:52:23 -07:00
Suresh Siddha
5d2bd7009f x86, fpu: decouple non-lazy/eager fpu restore from xsave
Decouple non-lazy/eager fpu restore policy from the existence of the xsave
feature. Introduce a synthetic CPUID flag to represent the eagerfpu
policy. "eagerfpu=on" boot paramter will enable the policy.

Requested-by: H. Peter Anvin <hpa@zytor.com>
Requested-by: Linus Torvalds <torvalds@linux-foundation.org>
Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1347300665-6209-2-git-send-email-suresh.b.siddha@intel.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:52:22 -07:00
Suresh Siddha
304bceda6a x86, fpu: use non-lazy fpu restore for processors supporting xsave
Fundamental model of the current Linux kernel is to lazily init and
restore FPU instead of restoring the task state during context switch.
This changes that fundamental lazy model to the non-lazy model for
the processors supporting xsave feature.

Reasons driving this model change are:

i. Newer processors support optimized state save/restore using xsaveopt and
xrstor by tracking the INIT state and MODIFIED state during context-switch.
This is faster than modifying the cr0.TS bit which has serializing semantics.

ii. Newer glibc versions use SSE for some of the optimized copy/clear routines.
With certain workloads (like boot, kernel-compilation etc), application
completes its work with in the first 5 task switches, thus taking upto 5 #DNA
traps with the kernel not getting a chance to apply the above mentioned
pre-load heuristic.

iii. Some xstate features (like AMD's LWP feature) don't honor the cr0.TS bit
and thus will not work correctly in the presence of lazy restore. Non-lazy
state restore is needed for enabling such features.

Some data on a two socket SNB system:
 * Saved 20K DNA exceptions during boot on a two socket SNB system.
 * Saved 50K DNA exceptions during kernel-compilation workload.
 * Improved throughput of the AVX based checksumming function inside the
   kernel by ~15% as xsave/xrstor is faster than the serializing clts/stts
   pair.

Also now kernel_fpu_begin/end() relies on the patched
alternative instructions. So move check_fpu() which uses the
kernel_fpu_begin/end() after alternative_instructions().

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1345842782-24175-7-git-send-email-suresh.b.siddha@intel.com
Merge 32-bit boot fix from,
Link: http://lkml.kernel.org/r/1347300665-6209-4-git-send-email-suresh.b.siddha@intel.com
Cc: Jim Kukunas <james.t.kukunas@linux.intel.com>
Cc: NeilBrown <neilb@suse.de>
Cc: Avi Kivity <avi@redhat.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:52:11 -07:00
Suresh Siddha
841e3604d3 x86, fpu: always use kernel_fpu_begin/end() for in-kernel FPU usage
use kernel_fpu_begin/end() instead of unconditionally accessing cr0 and
saving/restoring just the few used xmm/ymm registers.

This has some advantages like:
* If the task's FPU state is already active, then kernel_fpu_begin()
  will just save the user-state and avoiding the read/write of cr0.
  In general, cr0 accesses are much slower.

* Manual save/restore of xmm/ymm registers will affect the 'modified' and
  the 'init' optimizations brought in the by xsaveopt/xrstor
  infrastructure.

* Foward compatibility with future vector register extensions will be a
  problem if the xmm/ymm registers are manually saved and restored
  (corrupting the extended state of those vector registers).

With this patch, there was no significant difference in the xor throughput
using AVX, measured during boot.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1345842782-24175-5-git-send-email-suresh.b.siddha@intel.com
Cc: Jim Kukunas <james.t.kukunas@linux.intel.com>
Cc: NeilBrown <neilb@suse.de>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:52:08 -07:00
Suresh Siddha
9c1c3fac53 x86, kvm: use kernel_fpu_begin/end() in kvm_load/put_guest_fpu()
kvm's guest fpu save/restore should be wrapped around
kernel_fpu_begin/end(). This will avoid for example taking a DNA
in kvm_load_guest_fpu() when it tries to load the fpu immediately
after doing unlazy_fpu() on the host side.

More importantly this will prevent the host process fpu from being
corrupted.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1345842782-24175-4-git-send-email-suresh.b.siddha@intel.com
Cc: Avi Kivity <avi@redhat.com>
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:52:07 -07:00
Suresh Siddha
377ffbcc53 x86, fpu: remove unnecessary user_fpu_end() in save_xstate_sig()
Few lines below we do drop_fpu() which is more safer. Remove the
unnecessary user_fpu_end() in save_xstate_sig(), which allows
the drop_fpu() to ignore any pending exceptions from the user-space
and drop the current fpu.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1345842782-24175-3-git-send-email-suresh.b.siddha@intel.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:52:06 -07:00
Suresh Siddha
e962591749 x86, fpu: drop_fpu() before restoring new state from sigframe
No need to save the state with unlazy_fpu(), that is about to get overwritten
by the state from the signal frame. Instead use drop_fpu() and continue
to restore the new state.

Also fold the stop_fpu_preload() into drop_fpu().

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1345842782-24175-2-git-send-email-suresh.b.siddha@intel.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:52:05 -07:00
Suresh Siddha
72a671ced6 x86, fpu: Unify signal handling code paths for x86 and x86_64 kernels
Currently for x86 and x86_32 binaries, fpstate in the user sigframe is copied
to/from the fpstate in the task struct.

And in the case of signal delivery for x86_64 binaries, if the fpstate is live
in the CPU registers, then the live state is copied directly to the user
sigframe. Otherwise  fpstate in the task struct is copied to the user sigframe.
During restore, fpstate in the user sigframe is restored directly to the live
CPU registers.

Historically, different code paths led to different bugs. For example,
x86_64 code path was not preemption safe till recently. Also there is lot
of code duplication for support of new features like xsave etc.

Unify signal handling code paths for x86 and x86_64 kernels.

New strategy is as follows:

Signal delivery: Both for 32/64-bit frames, align the core math frame area to
64bytes as needed by xsave (this where the main fpu/extended state gets copied
to and excludes the legacy compatibility fsave header for the 32-bit [f]xsave
frames). If the state is live, copy the register state directly to the user
frame. If not live, copy the state in the thread struct to the user frame. And
for 32-bit [f]xsave frames, construct the fsave header separately before
the actual [f]xsave area.

Signal return: As the 32-bit frames with [f]xstate has an additional
'fsave' header, copy everything back from the user sigframe to the
fpstate in the task structure and reconstruct the fxstate from the 'fsave'
header (Also user passed pointers may not be correctly aligned for
any attempt to directly restore any partial state). At the next fpstate usage,
everything will be restored to the live CPU registers.
For all the 64-bit frames and the 32-bit fsave frame, restore the state from
the user sigframe directly to the live CPU registers. 64-bit signals always
restored the math frame directly, so we can expect the math frame pointer
to be correctly aligned. For 32-bit fsave frames, there are no alignment
requirements, so we can restore the state directly.

"lat_sig catch" microbenchmark numbers (for x86, x86_64, x86_32 binaries) are
with in the noise range with this change.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1343171129-2747-4-git-send-email-suresh.b.siddha@intel.com
[ Merged in compilation fix ]
Link: http://lkml.kernel.org/r/1344544736.8326.17.camel@sbsiddha-desk.sc.intel.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:51:48 -07:00
Suresh Siddha
0ca5bd0d88 x86, fpu: Consolidate inline asm routines for saving/restoring fpu state
Consolidate x86, x86_64 inline asm routines saving/restoring fpu state
using config_enabled().

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1343171129-2747-3-git-send-email-suresh.b.siddha@intel.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:51:26 -07:00
Suresh Siddha
050902c011 x86, signal: Cleanup ifdefs and is_ia32, is_x32
Use config_enabled() to cleanup the definitions of is_ia32/is_x32. Move
the function prototypes to the header file to cleanup ifdefs,
and move the x32_setup_rt_frame() code around.

Signed-off-by: Suresh Siddha <suresh.b.siddha@intel.com>
Link: http://lkml.kernel.org/r/1343171129-2747-2-git-send-email-suresh.b.siddha@intel.com
Merged in compilation fix from,
Link: http://lkml.kernel.org/r/1344544736.8326.17.camel@sbsiddha-desk.sc.intel.com
Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
2012-09-18 15:51:26 -07:00
Olof Johansson
23f3f0613f ARM i.MX: Two fixes for i.MX
- armadillo5x0 board broken since v3.5 (stable material)
 - i.MX25 Architecture broken since v3.6-rc1
 -----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1.4.12 (GNU/Linux)
 
 iQIcBAABCAAGBQJQWIlrAAoJEPFlmONMx+ezJTYP/3b7RbyCRfrYKebiytkmeaDc
 VDxsQekF3wANL8e0B2sZPY6TG8nagnAro2DvoJZJDMSx560JdIHEyMnzc11o0Gj7
 KE5XmiJFEyO/KNlZXVdv6LUohz97QNn475vjWcMQll1DYAXeRD3O5MstS23NBWc9
 HzNoVgG4wqlUPA+d4Y/+IMcnyRBgOzIQGO3ICoVp6GXJP9/9GXtUIaTWo4dQfMGB
 PT4W4c3QWsfHsuSbF5I53NVzMqMc7d86sY/WIu+qaUY1phlwB+yCBgpd0iHT9COI
 ph83LEur7COyeD2C7jCQJ5JNhD6qSqd8egOZXr/4CErTLAxqOsB3l736RoquBFbt
 L6PDo/SgeUvVW7Oeb2I3tzMML5UfvmHoewN7gUOgNlCn+cIBI38xb9STKmrBeJAV
 xxdxwr721d9Lt6ANmmkP5ZOO0AEJcoECRMnVamoXYjmIavg9kgVaktsPI4TaSm0f
 VyWt7qZGdjgdo/7wyZPcUoY67/E37n5yEATutl35ftjxIDkBFgHXU7VyZGSBYLhx
 vdteC7+islHf1b5jG1OlBDJT2YqKbZ57rmoYeJpAbZ/o2sDi2mjceuTPKSJkt5cd
 y+2CBZ3phQpMOgHVjEaurYZ5hDlGF3u6cM9yX8AxXrMk94TmgXoXzg/jvmzEVVuX
 rI78Y0Wto4EVsdOygVuw
 =2lcv
 -----END PGP SIGNATURE-----

Merge tag 'imx-fixes' of git://git.pengutronix.de/git/imx/linux-2.6 into fixes

From Sascha Hauer:

ARM i.MX: Two fixes for i.MX
- armadillo5x0 board broken since v3.5 (stable material)
- i.MX25 Architecture broken since v3.6-rc1

* tag 'imx-fixes' of git://git.pengutronix.de/git/imx/linux-2.6:
  ARM i.MX25: Make timer irq work again
  ARM: imx: armadillo5x0: Fix illegal register access
2012-09-18 10:16:44 -07:00
Sascha Hauer
42a3f891ca ARM i.MX25: Make timer irq work again
Since i.MX has SPARSE_IRQ enabled the i.MX25 timer is broken. This
is because the internal irqs now start at an offset of NR_IRQS_LEGACY.
The patch fixed this up, but missed the i.MX25 timer which used a
hardcoded value instead of a define. This patch introduces a define
for the timer irq and uses it.

This is broken since introduced with 3.6-rc1:

| commit 8842a9e286
| Author: Shawn Guo <shawn.guo@linaro.org>
| Date:   Thu Jun 14 11:16:14 2012 +0800
|
|    ARM: imx: enable SPARSE_IRQ for imx platform

Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Acked-by: Shawn Guo <shawn.guo@linaro.org>
2012-09-18 16:46:50 +02:00
Fabio Estevam
35495173e1 ARM: imx: armadillo5x0: Fix illegal register access
Since commit eb92044eb (ARM i.MX3: Make ccm base address a variable )
it is necessary to pass the CCM register base as a variable.

Fix the CCM register access in mach-armadillo5x0 by passing mx3_ccm_base and
avoid illegal accesses.

Also applies to v3.5

Reported-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Fabio Estevam <fabio.estevam@freescale.com>
Acked-by: Arnd Bergmann <arnd@arndb.de>
Signed-off-by: Sascha Hauer <s.hauer@pengutronix.de>
Cc: stable@vger.kernel.org
2012-09-18 16:46:24 +02:00
Olof Johansson
652f5b5e80 Modify AT91 device tree files for making the GPIO interrupts work.
-----BEGIN PGP SIGNATURE-----
 Version: GnuPG v1.4.11 (GNU/Linux)
 
 iQEcBAABAgAGBQJQWC9nAAoJEAf03oE53VmQV1QIAJ9zbpXxtPQytJjNR8Md9XEK
 Y5Ao2xzjZu4ZawQCyr3cnYvup7hMxfTK6Fij5lWwnwGech4tqfkoIuvL2soiLJfG
 4ioznzC3CN6rNZZTJo9RHIVouEADfqX6pShLPttau34RCDpZqiyuekcav8slBiCo
 /lOgsZWVOkjpjvvino4w0EdUjyz5jWT0IFcDxoVhnJGsdHunkXwASu8OeYk9kWSJ
 KbWERMeruGssz9ch6rccIvaYV7pL+WM2Rm8PDNHzvziKaXRD8Sod6qZJpj7mp9XJ
 9gAujzss7sZ65T1P3vVEqYbWFAZjYzUdFcj9G075kSlZysA8rthZVXfym+xg/G0=
 =7hqG
 -----END PGP SIGNATURE-----

Merge tag 'at91-fixes' of git://github.com/at91linux/linux-at91 into fixes

From Nicolas Ferre:

Modify AT91 device tree files for making the GPIO interrupts work.

* tag 'at91-fixes' of git://github.com/at91linux/linux-at91:
  ARM: at91: fix missing #interrupt-cells on gpio-controller
2012-09-18 07:43:16 -07:00
Olof Johansson
0921dcea79 Merge branch 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas into fixes
* 'fixes' of git://git.kernel.org/pub/scm/linux/kernel/git/horms/renesas:
  ARM: shmobile: kzm9g: bugfix: correct mmcif interrupt settings
2012-09-18 07:41:25 -07:00
Olof Johansson
d7235b8b92 Merge branch 'v3.6-samsung-fixes-3' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung into fixes
* 'v3.6-samsung-fixes-3' of git://git.kernel.org/pub/scm/linux/kernel/git/kgene/linux-samsung:
  ARM: SAMSUNG: Use spin_lock_{irqsave,irqrestore} in clk_set_rate
  ARM: SAMSUNG: use spin_lock_irqsave() in clk_set_parent
2012-09-18 07:40:02 -07:00
Al Viro
5e071e2b4b sh: Fix up TIF_NOTIFY_RESUME sans TIF_SIGPENDING handling.
As Al notes, we missed a TIF_NOTIFY_RESUME check which caused any
handlers without TIF_SIGPENDING also set to skip the notification:

	Looks like while it is in the relevant masks *and* checked in
	do_notify_resume() both on 32bit and 64bit variants since commit
	ab99c733ae ("sh: Make syscall tracer
	use tracehook notifiers, add TIF_NOTIFY_RESUME.") they are
	actually *not* reached without simulataneous SIGPENDING, since
	the actual glue in the callers had not been updated back then and
	still checks for _TIF_SIGPENDING alone when deciding whether to
	hit do_notify_resume() or not.

Reported-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj@renesas.com>
Signed-off-by: Al Viro <viro@zeniv.linux.org.uk>
Tested-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj@renesas.com>
Signed-off-by: Paul Mundt <lethal@linux-sh.org>
2012-09-18 17:04:37 +09:00
Tetsuyuki Kobayashi
a704835d6a ARM: shmobile: kzm9g: bugfix: correct mmcif interrupt settings
Correct interrupt settings of sh_mmc:int and sh_mmc:error in board-kzm9g.c.

Signed-off-by: Tetsuyuki Kobayashi <koba@kmckk.co.jp>
Acked-by: Kuninori Morimoto <kuninori.morimoto.gx@renesas.com>
Signed-off-by: Simon Horman <horms@verge.net.au>
2012-09-18 11:19:44 +09:00
Tushar Behera
d6838a62b4 ARM: SAMSUNG: Use spin_lock_{irqsave,irqrestore} in clk_set_rate
The spinlock clocks_lock can be held during ISR, hence it is not safe to
hold that lock with disabling interrupts.

It fixes following potential deadlock.

=========================================================
[ INFO: possible irq lock inversion dependency detected ]
3.6.0-rc4+ #2 Not tainted
---------------------------------------------------------
swapper/0/1 just changed the state of lock:
 (&(&host->lock)->rlock){-.....}, at: [<c027fb0d>] sdhci_irq+0x15/0x564
but this lock took another, HARDIRQ-unsafe lock in the past:
 (clocks_lock){+.+...}

and interrupts could create inverse lock ordering between them.

other info that might help us debug this:
 Possible interrupt unsafe locking scenario:

       CPU0                    CPU1
       ----                    ----
  lock(clocks_lock);
                               local_irq_disable();
                               lock(&(&host->lock)->rlock);
                               lock(clocks_lock);
  <Interrupt>
    lock(&(&host->lock)->rlock);

 *** DEADLOCK ***

Signed-off-by: Tushar Behera <tushar.behera@linaro.org>
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
2012-09-18 10:05:34 +09:00
Borislav Petkov
57639bedd2 x86, MCE: Remove unused defines
Those were sitting there unused since the dawn of time, drop them.

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
2012-09-17 19:33:38 +02:00
Borislav Petkov
e57dbaf77f x86, mce: Enable MCA support by default
MCA is the basic support for hardware error logging and reporting, and
it is majorly unwise to run without it so enable machine check software
support by default on x86.

Signed-off-by: Borislav Petkov <borislav.petkov@amd.com>
Acked-by: Tony Luck <tony.luck@intel.com>
2012-09-17 19:33:13 +02:00
Yan, Zheng
314d9f63f3 perf/x86: Add cpumask for uncore pmu
This patch adds a cpumask file to the uncore pmu sysfs directory.  The
cpumask file contains one active cpu for every socket.

Signed-off-by: "Yan, Zheng" <zheng.z.yan@intel.com>
Acked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
Acked-by: Ingo Molnar <mingo@kernel.org>
Cc: Andi Kleen <andi@firstfloor.org>
Cc: Jiri Olsa <jolsa@redhat.com>
Cc: Peter Zijlstra <a.p.zijlstra@chello.nl>
Cc: Stephane Eranian <eranian@google.com>
Cc: "Yan, Zheng" <zheng.z.yan@intel.com>
Link: http://lkml.kernel.org/r/1347263631-23175-2-git-send-email-zheng.z.yan@intel.com
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
2012-09-17 13:11:43 -03:00
Matthew Garrett
e9b10953ed x86, EFI: Calculate the EFI framebuffer size instead of trusting the firmware
Seth Forshee reported that his system was reporting that the EFI framebuffer
stretched from 0x90010000-0xb0010000 despite the GPU's BAR only covering
0x90000000-0x9ffffff. It's safer to calculate this value from the pixel
stride and screen height (values we already depend on) rather than face
potential problems with resource allocation later on.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Tested-by: Seth Forshee <seth.forshee@canonical.com>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>
2012-09-17 13:29:24 +01:00
Matthew Garrett
f462ed939d efifb: Skip DMI checks if the bootloader knows what it's doing
The majority of the DMI checks in efifb are for cases where the bootloader
has provided invalid information. However, on some machines the overrides
may do more harm than good due to configuration differences between machines
with the same machine identifier. It turns out that it's possible for the
bootloader to get the correct information on GOP-based systems, but we
can't guarantee that the kernel's being booted with one that's been updated
to do so. Add support for a capabilities flag that can be set by the
bootloader, and skip the DMI checks in that case. Additionally, set this
flag in the UEFI stub code.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Acked-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>
2012-09-17 13:29:23 +01:00
Seiji Aguchi
d6cf86d8f2 efi: initialize efi.runtime_version to make query_variable_info/update_capsule workable
A value of efi.runtime_version is checked before calling
update_capsule()/query_variable_info() as follows.
But it isn't initialized anywhere.

<snip>
static efi_status_t virt_efi_query_variable_info(u32 attr,
                                                 u64 *storage_space,
                                                 u64 *remaining_space,
                                                 u64 *max_variable_size)
{
        if (efi.runtime_version < EFI_2_00_SYSTEM_TABLE_REVISION)
                return EFI_UNSUPPORTED;
<snip>

This patch initializes a value of efi.runtime_version at boot time.

Signed-off-by: Seiji Aguchi <seiji.aguchi@hds.com>
Acked-by: Matthew Garrett <mjg@redhat.com>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>
2012-09-17 13:29:22 +01:00
Matthew Garrett
9dead5bbb8 efi: Build EFI stub with EFI-appropriate options
We can't assume the presence of the red zone while we're still in a boot
services environment, so we should build with -fno-red-zone to avoid
problems. Change the size of wchar at the same time to make string handling
simpler.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>
2012-09-17 13:29:21 +01:00
Matthew Garrett
38cb5ef447 X86: Improve GOP detection in the EFI boot stub
We currently use the PCI IO protocol as a proxy for a functional GOP. This
is less than ideal, since some platforms will put the GOP on output devices
rather than the GPU itself. Move to using the conout protocol. This is not
guaranteed per-spec, but is part of the consplitter implementation that
causes this problem in the first place and so should be reliable.

Signed-off-by: Matthew Garrett <mjg@redhat.com>
Signed-off-by: Matt Fleming <matt.fleming@intel.com>
2012-09-17 13:29:20 +01:00
Jean-Christophe PLAGNIOL-VILLARD
51ac51a6a5 ARM: at91: fix missing #interrupt-cells on gpio-controller
Signed-off-by: Jean-Christophe PLAGNIOL-VILLARD <plagnioj@jcrosoft.com>
Tested-by: Bo Shen <voice.shen@atmel.com>
Signed-off-by: Nicolas Ferre <nicolas.ferre@atmel.com>
2012-09-17 10:25:01 +02:00
Gerald Schaefer
4db84d4f07 s390/mm: fix user access page-table walk code
The s390 page-table walk code, used for user copy and futex, currently
cannot handle huge pages. As far as user copy is concerned, that is
not really a problem because those functions will only be used on old
hardware that has no huge page support. But the futex code will also
use pagetable walk functions on current hardware when user space runs
in primary space mode. So, if a futex sits in a huge page, the futex
operation on it will result in a page fault loop or even data
corruption.

This patch adds the code for resolving huge page mappings in the user
access pagetable walk code on s390.

Signed-off-by: Gerald Schaefer <gerald.schaefer@de.ibm.com>
Signed-off-by: Martin Schwidefsky <schwidefsky@de.ibm.com>
2012-09-17 09:44:24 +02:00
Mandeep Singh Baines
dbc5e1e89a ARM: SAMSUNG: use spin_lock_irqsave() in clk_set_parent
From 0cdf3aff, "ARM: SAMSUNG: use spin_lock_irqsave() in
clk_{enable,disable}":

  The clk_enable()and clk_disable() can be used process and ISR either.
  And actually it is used for real product and other platforms use it
  now. So spin_lock_irqsave() should be used instead.

We need to make a similar change in clk_set_parent(). Otherwise,
you can potentially get spinlock recursion:

BUG: spinlock recursion on CPU#0, kinteractive/68
 lock: 807832a8, .magic: dead4ead, .owner: kinteractive/68, .owner_cpu: 0
[<80015f54>] (unwind_backtrace+0x0/0x128) from [<804f2914>] (dump_stack+0x20/0x24)
[<804f2914>] (dump_stack+0x20/0x24) from [<804f57b8>] (spin_dump+0x80/0x94)
[<804f57b8>] (spin_dump+0x80/0x94) from [<804f57f8>] (spin_bug+0x2c/0x30)
[<804f57f8>] (spin_bug+0x2c/0x30) from [<80222730>] (do_raw_spin_lock+0x54/0x150)
[<80222730>] (do_raw_spin_lock+0x54/0x150) from [<804f96ec>] (_raw_spin_lock_irqsave+0x20/0x28)
[<804f96ec>] (_raw_spin_lock_irqsave+0x20/0x28) from [<80022ea4>] (clk_enable+0x3c/0x84)
[<80022ea4>] (clk_enable+0x3c/0x84) from [<8038336c>] (s5p_mfc_clock_on+0x60/0x74)
[<8038336c>] (s5p_mfc_clock_on+0x60/0x74) from [<8038645c>] (s5p_mfc_read_info+0x20/0x38)
[<8038645c>] (s5p_mfc_read_info+0x20/0x38) from [<8037ca3c>] (s5p_mfc_handle_frame+0x2e4/0x4bc)
[<8037ca3c>] (s5p_mfc_handle_frame+0x2e4/0x4bc) from [<8037d420>] (s5p_mfc_irq+0x1ec/0x6cc)
[<8037d420>] (s5p_mfc_irq+0x1ec/0x6cc) from [<8007fc74>] (handle_irq_event_percpu+0x8c/0x244)
[<8007fc74>] (handle_irq_event_percpu+0x8c/0x244) from [<8007fe78>] (handle_irq_event+0x4c/0x6c)
[<8007fe78>] (handle_irq_event+0x4c/0x6c) from [<80082dd8>] (handle_fasteoi_irq+0xe4/0x150)
[<80082dd8>] (handle_fasteoi_irq+0xe4/0x150) from [<8007f424>] (generic_handle_irq+0x3c/0x50)
[<8007f424>] (generic_handle_irq+0x3c/0x50) from [<8000f7c4>] (handle_IRQ+0x88/0xc8)
[<8000f7c4>] (handle_IRQ+0x88/0xc8) from [<80008564>] (gic_handle_irq+0x44/0x68)
[<80008564>] (gic_handle_irq+0x44/0x68) from [<8000e400>] (__irq_svc+0x40/0x60)
Exception stack(0xef3cbe68 to 0xef3cbeb0)
[<8000e400>] (__irq_svc+0x40/0x60) from [<80022cfc>] (clk_set_parent+0x30/0x74)
[<80022cfc>] (clk_set_parent+0x30/0x74) from [<803ac7f8>] (set_apll.isra.0+0x28/0xb0)
[<803ac7f8>] (set_apll.isra.0+0x28/0xb0) from [<803ac8e4>] (exynos5250_set_frequency+0x64/0xb8)
[<803ac8e4>] (exynos5250_set_frequency+0x64/0xb8) from [<803ac280>] (exynos_target+0x1b0/0x220)
[<803ac280>] (exynos_target+0x1b0/0x220) from [<803a4a0c>] (__cpufreq_driver_target+0xb0/0xd4)
[<803a4a0c>] (__cpufreq_driver_target+0xb0/0xd4) from [<803aab80>] (cpufreq_interactive_updown_task+0x214/0x264)
[<803aab80>] (cpufreq_interactive_updown_task+0x214/0x264) from [<80047d04>] (kthread+0x9c/0xa8)
[<80047d04>] (kthread+0x9c/0xa8) from [<8000fa48>] (kernel_thread_exit+0x0/0x8)

Signed-off-by: Mandeep Singh Baines <msb@chromium.org>
Suggested-by: Sunil Mazhavanchery <sunilm@samsung.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-samsung-soc@vger.kernel.org
Cc: Ben Dooks <ben-linux@fluff.org>
Cc: Russell King <linux@arm.linux.org.uk>
Cc: Minho Ban <mhban@samsung.com>
Cc: Jaecheol Lee <jc.lee@samsung.com>
Cc: Sunyoung Kang <sy0816.kang@samsung.com>
Cc: Olof Johansson <olofj@chromium.org>
Signed-off-by: Kukjin Kim <kgene.kim@samsung.com>
2012-09-17 13:13:42 +09:00
Ralf Baechle
e376fdf420 MIPS: Malta: Don't crash on spurious interrupt.
48d480b0bd [[MIPS] Malta: Fix off by one bug in interrupt
handler.] did not take in account that irq_ffs() will also return 0 if for some reason
the set of pending interrupts happens to be empty.

This is trivial to trigger with a RM5261 CPU module running a 64-bit kernel and results
in something like the following:

CPU 0 Unable to handle kernel paging request at virtual address 0000000000000000, epc == ffffffff801772d0, ra == ffffffff8017ad24
Oops[#1]:
Cpu 0
$ 0   : 0000000000000000 ffffffff9000a4e0 ffffffff9000a4e0 ffffffff9000a4e0
$ 4   : ffffffff80592be0 0000000000000000 00000000000000d6 ffffffff80322ed0
$ 8   : ffffffff805fe538 0000000000000000 ffffffff9000a4e0 ffffffff80590000
$12   : 00000000000000d6 0000000000000000 ffffffff80600000 ffffffff805fe538
$16   : 0000000000000000 0000000000000010 ffffffff80592be0 0000000000000010
$20   : 0000000000000000 0000000000500001 0000000000000000 ffffffff8051e078
$24   : 0000000000000028 ffffffff803226e8
$28   : 9800000003828000 980000000382b900 ffffffff8051e060 ffffffff8017ad24
Hi    : 0000000000000000
Lo    : 0000006388974000
epc   : ffffffff801772d0 handle_irq_event_percpu+0x70/0x2f0
    Not tainted
ra    : ffffffff8017ad24 handle_percpu_irq+0x54/0x88
Status: 9000a4e2    KX SX UX KERNEL EXL
Cause : 00808008
BadVA : 0000000000000000
PrId  : 000028a0 (Nevada)
Modules linked in:
Process init (pid: 1, threadinfo=9800000003828000, task=9800000003827968, tls=0000000077087490)
Stack : ffffffff80592be0 ffffffff8058d248 0000000000000040 0000000000000000
        ffffffff80613340 0000000000500001 ffffffff805a0000 0000000000000882
        9800000003b89000 ffffffff8017ad24 00000000000000d5 0000000000000010
        ffffffff9000a4e1 ffffffff801769f4 ffffffff9000a4e0 ffffffff801037f8
        0000000000000000 ffffffff80101c44 0000000000000000 ffffffff9000a4e0
        0000000000000000 9000000018000000 90000000180003f9 0000000000000001
        0000000000000000 00000000000000ff 0000000000000018 0000000000000001
        0000000000000001 00000000003fffff 0000000000000020 ffffffff802cf7ac
        ffffffff80208918 000000007fdadf08 ffffffff80612d88 ffffffff9000a4e1
        0000000000000040 0000000000000000 ffffffff80613340 0000000000500001
        ...
Call Trace:
[<ffffffff801772d0>] handle_irq_event_percpu+0x70/0x2f0
[<ffffffff8017ad24>] handle_percpu_irq+0x54/0x88
[<ffffffff801769f4>] generic_handle_irq+0x44/0x60
[<ffffffff801037f8>] do_IRQ+0x48/0x70
[<ffffffff80101c44>] ret_from_irq+0x0/0x4
[<ffffffff80326170>] serial8250_startup+0x310/0x870
[<ffffffff8032175c>] uart_startup.part.7+0x9c/0x330
[<ffffffff80321b4c>] uart_open+0x15c/0x1b0
[<ffffffff80302034>] tty_open+0x1fc/0x720
[<ffffffff801bffac>] chrdev_open+0x7c/0x180
[<ffffffff801b9ab8>] do_dentry_open.isra.14+0x288/0x390
[<ffffffff801bac5c>] nameidata_to_filp+0x5c/0xc0
[<ffffffff801ca700>] do_last.isra.33+0x330/0x8f0
[<ffffffff801caf3c>] path_openat+0xbc/0x440
[<ffffffff801cb3c8>] do_filp_open+0x38/0xa8
[<ffffffff801bade4>] do_sys_open+0x124/0x218
[<ffffffff80110538>] handle_sys+0x118/0x13c

Code: 02d5a825  12800012  02a0b02d <de820000> de850008  0040f809  0220202d  0040a82d  40026000
---[ end trace 5d8e7b9a86badd2d ]---
Kernel panic - not syncing: Fatal exception in interrupt

Signed-off-by: Ralf Baechle <ralf@linux-mips.org>
2012-09-17 02:04:29 +02:00
Matthew Leach
e1e5b7e425 ARM: 7532/1: decompressor: reset SCTLR.TRE for VMSA ARMv7 cores
This patch zeroes the SCTLR.TRE bit prior to setting the mapping as
cacheable for ARMv7 cores in the decompressor, ensuring that the
memory region attributes are obtained from the C and B bits, not from
the page tables.

Cc: Nicolas Pitre <nico@fluxnic.net>
Reviewed-by: Will Deacon <will.deacon@arm.com>
Signed-off-by: Matthew Leach <matthew.leach@arm.com>
Signed-off-by: Will Deacon <will.deacon@arm.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2012-09-16 00:16:16 +01:00
Lars-Peter Clausen
8ef997b67f ARM: 7534/1: clk: Make the managed clk functions generically available
The managed clk functions are currently only available when the generic clk
lookup framework is build. But the managed clk functions are merely wrappers
around clk_get and clk_put and do not depend on any specifics of the generic
lookup functions and there are still quite a few custom implementations of the
clk API. So make the managed functions available whenever the clk API is
implemented.

The patch also removes the custom implementation of devm_clk_get for the
coldfire platform.

Signed-off-by: Lars-Peter Clausen <lars@metafoo.de>
Acked-by: Greg Ungerer <gerg@uclinux.org>
Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2012-09-15 21:55:27 +01:00
Oleg Nesterov
baedbf02b1 uprobes: Make arch_uprobe_task->saved_trap_nr "unsigned int"
Make arch_uprobe_task->saved_trap_nr "unsigned int" and move it down
after ->saved_scratch_register, this changes sizeof() from 24 to 16.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
2012-09-15 17:37:32 +02:00
Oleg Nesterov
d6a00b35e4 uprobes/x86: Fix arch_uprobe_disable_step() && UTASK_SSTEP_TRAPPED interaction
arch_uprobe_disable_step() should also take UTASK_SSTEP_TRAPPED into
account. In this case the probed insn was not executed, we need to
clear X86_EFLAGS_TF if it was set by us and that is all.

Again, this code will look more clean when we move it into
arch_uprobe_post_xol() and arch_uprobe_abort_xol().

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
2012-09-15 17:37:32 +02:00
Oleg Nesterov
3a4664aa83 uprobes/x86: Xol should send SIGTRAP if X86_EFLAGS_TF was set
arch_uprobe_disable_step() correctly preserves X86_EFLAGS_TF and
returns to user-mode. But this means the application gets SIGTRAP
only after the next insn.

This means that UPROBE_CLEAR_TF logic is not really right. _enable
should only record the state of X86_EFLAGS_TF, and _disable should
check it separately from UPROBE_FIX_SETF.

Remove arch_uprobe_task->restore_flags, add ->saved_tf instead, and
change enable/disable accordingly. This assumes that the probed insn
was not trapped, see the next patch.

arch_uprobe_skip_sstep() logic has the same problem, change it to
check X86_EFLAGS_TF and send SIGTRAP as well. We will cleanup this
all after we fold enable/disable_step into pre/post_hol hooks.

Note: send_sig(SIGTRAP) is not actually right, we need send_sigtrap().
But this needs more changes, handle_swbp() does the same and this is
equally wrong.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
2012-09-15 17:37:31 +02:00
Oleg Nesterov
9bd1190a11 uprobes/x86: Do not (ab)use TIF_SINGLESTEP/user_*_single_step() for single-stepping
user_enable/disable_single_step() was designed for ptrace, it assumes
a single user and does unnecessary and wrong things for uprobes. For
example:

	- arch_uprobe_enable_step() can't trust TIF_SINGLESTEP, an
	  application itself can set X86_EFLAGS_TF which must be
	  preserved after arch_uprobe_disable_step().

	- we do not want to set TIF_SINGLESTEP/TIF_FORCED_TF in
	  arch_uprobe_enable_step(), this only makes sense for ptrace.

	- otoh we leak TIF_SINGLESTEP if arch_uprobe_disable_step()
	  doesn't do user_disable_single_step(), the application will
	  be killed after the next syscall.

	- arch_uprobe_enable_step() does access_process_vm() we do
	  not need/want.

Change arch_uprobe_enable/disable_step() to set/clear X86_EFLAGS_TF
directly, this is much simpler and more correct. However, we need to
clear TIF_BLOCKSTEP/DEBUGCTLMSR_BTF before executing the probed insn,
add set_task_blockstep(false).

Note: with or without this patch, there is another (hopefully minor)
problem. A probed "pushf" insn can see the wrong X86_EFLAGS_TF set by
uprobes. Perhaps we should change _disable to update the stack, or
teach arch_uprobe_skip_sstep() to emulate this insn.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
Acked-by: Srikar Dronamraju <srikar@linux.vnet.ibm.com>
2012-09-15 17:37:30 +02:00
Oleg Nesterov
95cf00fa5d ptrace/x86: Partly fix set_task_blockstep()->update_debugctlmsr() logic
Afaics the usage of update_debugctlmsr() and TIF_BLOCKSTEP in
step.c was always very wrong.

1. update_debugctlmsr() was simply unneeded. The child sleeps
   TASK_TRACED, __switch_to_xtra(next_p => child) should notice
   TIF_BLOCKSTEP and set/clear DEBUGCTLMSR_BTF after resume if
   needed.

2. It is wrong. The state of DEBUGCTLMSR_BTF bit in CPU register
   should always match the state of current's TIF_BLOCKSTEP bit.

3. Even get_debugctlmsr() + update_debugctlmsr() itself does not
   look right. Irq can change other bits in MSR_IA32_DEBUGCTLMSR
   register or the caller can be preempted in between.

4. It is not safe to play with TIF_BLOCKSTEP if task != current.
   DEBUGCTLMSR_BTF and TIF_BLOCKSTEP should always match each
   other if the task is running. The tracee is stopped but it
   can be SIGKILL'ed right before set/clear_tsk_thread_flag().

However, now that uprobes uses user_enable_single_step(current)
we can't simply remove update_debugctlmsr(). So this patch adds
the additional "task == current" check and disables irqs to avoid
the race with interrupts/preemption.

Unfortunately this patch doesn't solve the last problem, we need
another fix. Probably we should teach ptrace_stop() to set/clear
single/block stepping after resume.

And afaics there is yet another problem: perf can play with
MSR_IA32_DEBUGCTLMSR from nmi, this obviously means that even
__switch_to_xtra() has problems.

Signed-off-by: Oleg Nesterov <oleg@redhat.com>
2012-09-15 17:37:29 +02:00