linux-toradex.git/arch/x86/include, branch v4.10

x86/CPU/AMD: Bring back Compute Unit ID

2017-02-05T11:18:45+00:00

Commit:

  a33d331761bc ("x86/CPU/AMD: Fix Bulldozer topology")

restored the initial approach we had with the Fam15h topology of
enumerating CU (Compute Unit) threads as cores. And this is still
correct - they're beefier than HT threads but still have some
shared functionality.

Our current approach has a problem with the Mad Max Steam game, for
example. Yves Dionne reported a certain "choppiness" while playing on
v4.9.5.

That problem stems most likely from the fact that the CU threads share
resources within one CU and when we schedule to a thread of a different
compute unit, this incurs latency due to migrating the working set to a
different CU through the caches.

When the thread siblings mask mirrors that aspect of the CUs and
threads, the scheduler pays attention to it and tries to schedule within
one CU first. Which takes care of the latency, of course.

Reported-by: Yves Dionne 
Signed-off-by: Borislav Petkov 
Cc:  # 4.9
Cc: Brice Goglin 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: Yazen Ghannam 
Link: http://lkml.kernel.org/r/20170205105022.8705-1-bp@alien8.de
Signed-off-by: Ingo Molnar

x86/microcode: Do not access the initrd after it has been freed

2017-01-30T08:32:42+00:00

When we look for microcode blobs, we first try builtin and if that
doesn't succeed, we fallback to the initrd supplied to the kernel.

However, at some point doing boot, that initrd gets jettisoned and we
shouldn't access it anymore. But we do, as the below KASAN report shows.
That's because find_microcode_in_initrd() doesn't check whether the
initrd is still valid or not.

So do that.

  ==================================================================
  BUG: KASAN: use-after-free in find_cpio_data
  Read of size 1 by task swapper/1/0
  page:ffffea0000db9d40 count:0 mapcount:0 mapping:          (null) index:0x1
  flags: 0x100000000000000()
  raw: 0100000000000000 0000000000000000 0000000000000001 00000000ffffffff
  raw: dead000000000100 dead000000000200 0000000000000000 0000000000000000
  page dumped because: kasan: bad access detected
  CPU: 1 PID: 0 Comm: swapper/1 Tainted: G        W       4.10.0-rc5-debug-00075-g2dbde22 #3
  Hardware name: Dell Inc. XPS 13 9360/0839Y6, BIOS 1.2.3 12/01/2016
  Call Trace:
   dump_stack
   ? _atomic_dec_and_lock
   ? __dump_page
   kasan_report_error
   ? pointer
   ? find_cpio_data
   __asan_report_load1_noabort
   ? find_cpio_data
   find_cpio_data
   ? vsprintf
   ? dump_stack
   ? get_ucode_user
   ? print_usage_bug
   find_microcode_in_initrd
   __load_ucode_intel
   ? collect_cpu_info_early
   ? debug_check_no_locks_freed
   load_ucode_intel_ap
   ? collect_cpu_info
   ? trace_hardirqs_on
   ? flat_send_IPI_mask_allbutself
   load_ucode_ap
   ? get_builtin_firmware
   ? flush_tlb_func
   ? do_raw_spin_trylock
   ? cpumask_weight
   cpu_init
   ? trace_hardirqs_off
   ? play_dead_common
   ? native_play_dead
   ? hlt_play_dead
   ? syscall_init
   ? arch_cpu_idle_dead
   ? do_idle
   start_secondary
   start_cpu
  Memory state around the buggy address:
   ffff880036e74f00: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
   ffff880036e74f80: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
  >ffff880036e75000: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
                     ^
   ffff880036e75080: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
   ffff880036e75100: ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff ff
  ==================================================================

Reported-by: Andrey Ryabinin 
Tested-by: Andrey Ryabinin 
Signed-off-by: Borislav Petkov 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: http://lkml.kernel.org/r/20170126165833.evjemhbqzaepirxo@pd.tnic
Signed-off-by: Ingo Molnar

x86/unwind: Include __schedule() in stack traces

2017-01-12T08:28:28+00:00

In the following commit:

  0100301bfdf5 ("sched/x86: Rewrite the switch_to() code")

... the layout of the 'inactive_task_frame' struct was designed to have
a frame pointer header embedded in it, so that the unwinder could use
the 'bp' and 'ret_addr' fields to report __schedule() on the stack (or
ret_from_fork() for newly forked tasks which haven't actually run yet).

Finish the job by changing get_frame_pointer() to return a pointer to
inactive_task_frame's 'bp' field rather than 'bp' itself.  This allows
the unwinder to start one frame higher on the stack, so that it properly
reports __schedule().

Reported-by: Miroslav Benes 
Signed-off-by: Josh Poimboeuf 
Cc: Andy Lutomirski 
Cc: Andy Lutomirski 
Cc: Borislav Petkov 
Cc: Brian Gerst 
Cc: Dave Jones 
Cc: Denys Vlasenko 
Cc: Dmitry Vyukov 
Cc: H. Peter Anvin 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: http://lkml.kernel.org/r/598e9f7505ed0aba86e8b9590aa528c6c7ae8dcd.1483978430.git.jpoimboe@redhat.com
Signed-off-by: Ingo Molnar

x86/unwind: Disable KASAN checks for non-current tasks

2017-01-12T08:28:27+00:00

There are a handful of callers to save_stack_trace_tsk() and
show_stack() which try to unwind the stack of a task other than current.
In such cases, it's remotely possible that the task is running on one
CPU while the unwinder is reading its stack from another CPU, causing
the unwinder to see stack corruption.

These cases seem to be mostly harmless.  The unwinder has checks which
prevent it from following bad pointers beyond the bounds of the stack.
So it's not really a bug as long as the caller understands that
unwinding another task will not always succeed.

In such cases, it's possible that the unwinder may read a KASAN-poisoned
region of the stack.  Account for that by using READ_ONCE_NOCHECK() when
reading the stack of another task.

Use READ_ONCE() when reading the stack of the current task, since KASAN
warnings can still be useful for finding bugs in that case.

Reported-by: Dmitry Vyukov 
Signed-off-by: Josh Poimboeuf 
Cc: Andy Lutomirski 
Cc: Andy Lutomirski 
Cc: Borislav Petkov 
Cc: Brian Gerst 
Cc: Dave Jones 
Cc: Denys Vlasenko 
Cc: H. Peter Anvin 
Cc: Linus Torvalds 
Cc: Miroslav Benes 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: http://lkml.kernel.org/r/4c575eb288ba9f73d498dfe0acde2f58674598f1.1483978430.git.jpoimboe@redhat.com
Signed-off-by: Ingo Molnar

x86/microcode/intel: Add a helper which gives the microcode revision

2017-01-09T22:11:14+00:00

Since on Intel we're required to do CPUID(1) first, before reading
the microcode revision MSR, let's add a special helper which does the
required steps so that we don't forget to do them next time, when we
want to read the microcode revision.

Signed-off-by: Borislav Petkov 
Link: http://lkml.kernel.org/r/20170109114147.5082-4-bp@alien8.de
Signed-off-by: Thomas Gleixner

x86/CPU: Add native CPUID variants returning a single datum

2017-01-09T22:11:13+00:00

... similarly to the cpuid_() variants.

Signed-off-by: Borislav Petkov 
Link: http://lkml.kernel.org/r/20170109114147.5082-2-bp@alien8.de
Signed-off-by: Thomas Gleixner

x86/cpu: Fix typo in the comment for Anniedale

2017-01-05T08:03:29+00:00

The proper spelling of Anniedale SoC with 'e' in the middle. Fix typo in the
comment line in intel-family.h header.

Signed-off-by: Andy Shevchenko 
Cc: Dave Hansen 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: http://lkml.kernel.org/r/20170102092229.87036-1-andriy.shevchenko@linux.intel.com
Signed-off-by: Ingo Molnar

mm: optimize PageWaiters bit use for unlock_page()

2016-12-29T19:03:15+00:00

In commit 62906027091f ("mm: add PageWaiters indicating tasks are
waiting for a page bit") Nick Piggin made our page locking no longer
unconditionally touch the hashed page waitqueue, which not only helps
performance in general, but is particularly helpful on NUMA machines
where the hashed wait queues can bounce around a lot.

However, the "clear lock bit atomically and then test the waiters bit"
sequence turns out to be much more expensive than it needs to be,
because you get a nasty stall when trying to access the same word that
just got updated atomically.

On architectures where locking is done with LL/SC, this would be trivial
to fix with a new primitive that clears one bit and tests another
atomically, but that ends up not working on x86, where the only atomic
operations that return the result end up being cmpxchg and xadd.  The
atomic bit operations return the old value of the same bit we changed,
not the value of an unrelated bit.

On x86, we could put the lock bit in the high bit of the byte, and use
"xadd" with that bit (where the overflow ends up not touching other
bits), and look at the other bits of the result.  However, an even
simpler model is to just use a regular atomic "and" to clear the lock
bit, and then the sign bit in eflags will indicate the resulting state
of the unrelated bit #7.

So by moving the PageWaiters bit up to bit #7, we can atomically clear
the lock bit and test the waiters bit on x86 too.  And architectures
with LL/SC (which is all the usual RISC suspects), the particular bit
doesn't matter, so they are fine with this approach too.

This avoids the extra access to the same atomic word, and thus avoids
the costly stall at page unlock time.

The only downside is that the interface ends up being a bit odd and
specialized: clear a bit in a byte, and test the sign bit.  Nick doesn't
love the resulting name of the new primitive, but I'd rather make the
name be descriptive and very clear about the limitation imposed by
trying to work across all relevant architectures than make it be some
generic thing that doesn't make the odd semantics explicit.

So this introduces the new architecture primitive

    clear_bit_unlock_is_negative_byte();

and adds the trivial implementation for x86.  We have a generic
non-optimized fallback (that just does a "clear_bit()"+"test_bit(7)"
combination) which can be overridden by any architecture that can do
better.  According to Nick, Power has the same hickup x86 has, for
example, but some other architectures may not even care.

All these optimizations mean that my page locking stress-test (which is
just executing a lot of small short-lived shell scripts: "make test" in
the git source tree) no longer makes our page locking look horribly bad.
Before all these optimizations, just the unlock_page() costs were just
over 3% of all CPU overhead on "make test".  After this, it's down to
0.66%, so just a quarter of the cost it used to be.

(The difference on NUMA is bigger, but there this micro-optimization is
likely less noticeable, since the big issue on NUMA was not the accesses
to 'struct page', but the waitqueue accesses that were already removed
by Nick's earlier commit).

Acked-by: Nick Piggin 
Cc: Dave Hansen 
Cc: Bob Peterson 
Cc: Steven Whitehouse 
Cc: Andrew Lutomirski 
Cc: Andreas Gruenbacher 
Cc: Peter Zijlstra 
Cc: Mel Gorman 
Signed-off-by: Linus Torvalds

clocksource: Use a plain u64 instead of cycle_t

2016-12-25T10:04:12+00:00

There is no point in having an extra type for extra confusion. u64 is
unambiguous.

Conversion was done with the following coccinelle script:

@rem@
@@
-typedef u64 cycle_t;

@fix@
typedef cycle_t;
@@
-cycle_t
+u64

Signed-off-by: Thomas Gleixner 
Cc: Peter Zijlstra 
Cc: John Stultz

Replace with globally

2016-12-24T19:46:01+00:00

This was entirely automated, using the script by Al:

  PATT='^[[:blank:]]*#[[:blank:]]*include[[:blank:]]*'
  sed -i -e "s!$PATT!#include !" \
        $(git grep -l "$PATT"|grep -v ^include/linux/uaccess.h)

to do the replacement at the end of the merge window.

Requested-by: Al Viro 
Signed-off-by: Linus Torvalds