linux-toradex.git/arch/x86/power, branch v4.9.11

x86/asm: Get rid of __read_cr4_safe()

2016-09-30T10:40:12+00:00

We use __read_cr4() vs __read_cr4_safe() inconsistently.  On
CR4-less CPUs, all CR4 bits are effectively clear, so we can make
the code simpler and more robust by making __read_cr4() always fix
up faults on 32-bit kernels.

This may fix some bugs on old 486-like CPUs, but I don't have any
easy way to test that.

Signed-off-by: Andy Lutomirski 
Cc: Brian Gerst 
Cc: Borislav Petkov 
Cc: david@saggiorato.net
Link: http://lkml.kernel.org/r/ea647033d357d9ce2ad2bbde5a631045f5052fb6.1475178370.git.luto@kernel.org
Signed-off-by: Thomas Gleixner

x86/power/64: Use __pa() for physical address computation

2016-08-15T22:39:37+00:00

The value of temp_level4_pgt is the physical address of the
top-level page directory, so use __pa() to compute it.

Signed-off-by: Rafael J. Wysocki 
Acked-by: Ingo Molnar

x86/power/64: Always create temporary identity mapping correctly

2016-08-08T20:04:30+00:00

The low-level resume-from-hibernation code on x86-64 uses
kernel_ident_mapping_init() to create the temoprary identity mapping,
but that function assumes that the offset between kernel virtual
addresses and physical addresses is aligned on the PGD level.

However, with a randomized identity mapping base, it may be aligned
on the PUD level and if that happens, the temporary identity mapping
created by set_up_temporary_mappings() will not reflect the actual
kernel identity mapping and the image restoration will fail as a
result (leading to a kernel panic most of the time).

To fix this problem, rework kernel_ident_mapping_init() to support
unaligned offsets between KVA and PA up to the PMD level and make
set_up_temporary_mappings() use it as approprtiate.

Reported-and-tested-by: Thomas Garnier 
Reported-by: Borislav Petkov 
Suggested-by: Yinghai Lu 
Signed-off-by: Rafael J. Wysocki 
Acked-by: Yinghai Lu

x86/power/64: Do not refer to __PAGE_OFFSET from assembly code

2016-08-02T23:35:38+00:00

When CONFIG_RANDOMIZE_MEMORY is set on x86-64, __PAGE_OFFSET becomes
a variable and using it as a symbol in the image memory restoration
assembly code under core_restore_code is not correct any more.

To avoid that problem, modify set_up_temporary_mappings() to compute
the physical address of the temporary page tables and store it in
temp_level4_pgt, so that the value of that variable is ready to be
written into CR3.  Then, the assembly code doesn't have to worry
about converting that value into a physical address and things work
regardless of whether or not CONFIG_RANDOMIZE_MEMORY is set.

Reported-and-tested-by: Thomas Garnier 
Signed-off-by: Rafael J. Wysocki

x86/power/64: Fix hibernation return address corruption

2016-07-29T11:38:59+00:00

In kernel bug 150021, a kernel panic was reported when restoring a
hibernate image.  Only a picture of the oops was reported, so I can't
paste the whole thing here.  But here are the most interesting parts:

  kernel tried to execute NX-protected page - exploit attempt? (uid: 0)
  BUG: unable to handle kernel paging request at ffff8804615cfd78
  ...
  RIP: ffff8804615cfd78
  RSP: ffff8804615f0000
  RBP: ffff8804615cfdc0
  ...
  Call Trace:
   do_signal+0x23
   exit_to_usermode_loop+0x64
   ...

The RIP is on the same page as RBP, so it apparently started executing
on the stack.

The bug was bisected to commit ef0f3ed5a4ac (x86/asm/power: Create
stack frames in hibernate_asm_64.S), which in retrospect seems quite
dangerous, since that code saves and restores the stack pointer from a
global variable ('saved_context').

There are a lot of moving parts in the hibernate save and restore paths,
so I don't know exactly what caused the panic.  Presumably, a FRAME_END
was executed without the corresponding FRAME_BEGIN, or vice versa.  That
would corrupt the return address on the stack and would be consistent
with the details of the above panic.

[ rjw: One major problem is that by the time the FRAME_BEGIN in
  restore_registers() is executed, the stack pointer value may not
  be valid any more.  Namely, the stack area pointed to by it
  previously may have been overwritten by some image memory contents
  and that page frame may now be used for whatever different purpose
  it had been allocated for before hibernation.  In that case, the
  FRAME_BEGIN will corrupt that memory. ]

Instead of doing the frame pointer save/restore around the bounds of the
affected functions, just do it around the call to swsusp_save().

That has the same effect of ensuring that if swsusp_save() sleeps, the
frame pointers will be correct.  It's also a much more obviously safe
way to do it than the original patch.  And objtool still doesn't report
any warnings.

Fixes: ef0f3ed5a4ac (x86/asm/power: Create stack frames in hibernate_asm_64.S)
Link: https://bugzilla.kernel.org/show_bug.cgi?id=150021
Cc: 4.6+  # 4.6+
Reported-by: Andre Reinke 
Tested-by: Andre Reinke 
Signed-off-by: Josh Poimboeuf 
Acked-by: Ingo Molnar 
Signed-off-by: Rafael J. Wysocki

x86 / hibernate: Use hlt_play_dead() when resuming from hibernation

2016-07-15T20:42:48+00:00

On Intel hardware, native_play_dead() uses mwait_play_dead() by
default and only falls back to the other methods if that fails.
That also happens during resume from hibernation, when the restore
(boot) kernel runs disable_nonboot_cpus() to take all of the CPUs
except for the boot one offline.

However, that is problematic, because the address passed to
__monitor() in mwait_play_dead() is likely to be written to in the
last phase of hibernate image restoration and that causes the "dead"
CPU to start executing instructions again.  Unfortunately, the page
containing the address in that CPU's instruction pointer may not be
valid any more at that point.

First, that page may have been overwritten with image kernel memory
contents already, so the instructions the CPU attempts to execute may
simply be invalid.  Second, the page tables previously used by that
CPU may have been overwritten by image kernel memory contents, so the
address in its instruction pointer is impossible to resolve then.

A report from Varun Koyyalagunta and investigation carried out by
Chen Yu show that the latter sometimes happens in practice.

To prevent it from happening, temporarily change the smp_ops.play_dead
pointer during resume from hibernation so that it points to a special
"play dead" routine which uses hlt_play_dead() and avoids the
inadvertent "revivals" of "dead" CPUs this way.

A slightly unpleasant consequence of this change is that if the
system is hibernated with one or more CPUs offline, it will generally
draw more power after resume than it did before hibernation, because
the physical state entered by CPUs via hlt_play_dead() is higher-power
than the mwait_play_dead() one in the majority of cases.  It is
possible to work around this, but it is unclear how much of a problem
that's going to be in practice, so the workaround will be implemented
later if it turns out to be necessary.

Link: https://bugzilla.kernel.org/show_bug.cgi?id=106371
Reported-by: Varun Koyyalagunta 
Original-by: Chen Yu 
Tested-by: Chen Yu 
Signed-off-by: Rafael J. Wysocki 
Acked-by: Ingo Molnar

x86/power/64: Fix kernel text mapping corruption during image restoration

2016-06-30T21:57:15+00:00

Logan Gunthorpe reports that hibernation stopped working reliably for
him after commit ab76f7b4ab23 (x86/mm: Set NX on gap between __ex_table
and rodata).

That turns out to be a consequence of a long-standing issue with the
64-bit image restoration code on x86, which is that the temporary
page tables set up by it to avoid page tables corruption when the
last bits of the image kernel's memory contents are copied into
their original page frames re-use the boot kernel's text mapping,
but that mapping may very well get corrupted just like any other
part of the page tables.  Of course, if that happens, the final
jump to the image kernel's entry point will go to nowhere.

The exact reason why commit ab76f7b4ab23 matters here is that it
sometimes causes a PMD of a large page to be split into PTEs
that are allocated dynamically and get corrupted during image
restoration as described above.

To fix that issue note that the code copying the last bits of the
image kernel's memory contents to the page frames occupied by them
previoulsy doesn't use the kernel text mapping, because it runs from
a special page covered by the identity mapping set up for that code
from scratch.  Hence, the kernel text mapping is only needed before
that code starts to run and then it will only be used just for the
final jump to the image kernel's entry point.

Accordingly, the temporary page tables set up in swsusp_arch_resume()
on x86-64 need to contain the kernel text mapping too.  That mapping
is only going to be used for the final jump to the image kernel, so
it only needs to cover the image kernel's entry point, because the
first thing the image kernel does after getting control back is to
switch over to its own original page tables.  Moreover, the virtual
address of the image kernel's entry point in that mapping has to be
the same as the one mapped by the image kernel's page tables.

With that in mind, modify the x86-64's arch_hibernation_header_save()
and arch_hibernation_header_restore() routines to pass the physical
address of the image kernel's entry point (in addition to its virtual
address) to the boot kernel (a small piece of assembly code involved
in passing the entry point's virtual address to the image kernel is
not necessary any more after that, so drop it).  Update RESTORE_MAGIC
too to reflect the image header format change.

Next, in set_up_temporary_mappings(), use the physical and virtual
addresses of the image kernel's entry point passed in the image
header to set up a minimum kernel text mapping (using memory pages
that won't be overwritten by the image kernel's memory contents) that
will map those addresses to each other as appropriate.

This makes the concern about the possible corruption of the original
boot kernel text mapping go away and if the the minimum kernel text
mapping used for the final jump marks the image kernel's entry point
memory as executable, the jump to it is guaraneed to succeed.

Fixes: ab76f7b4ab23 (x86/mm: Set NX on gap between __ex_table and rodata)
Link: http://marc.info/?l=linux-pm&m=146372852823760&w=2
Reported-by: Logan Gunthorpe 
Reported-and-tested-by: Borislav Petkov 
Tested-by: Kees Cook 
Signed-off-by: Rafael J. Wysocki

x86/cpufeature: Remove cpu_has_pse

2016-03-31T11:35:10+00:00

Signed-off-by: Borislav Petkov 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Link: http://lkml.kernel.org/r/1459266123-21878-11-git-send-email-bp@alien8.de
Signed-off-by: Ingo Molnar

x86/asm/power: Create stack frames in hibernate_asm_64.S

2016-02-24T07:35:43+00:00

swsusp_arch_suspend() and restore_registers() are callable non-leaf
functions which don't honor CONFIG_FRAME_POINTER, which can result in
bad stack traces.  Also they aren't annotated as ELF callable functions
which can confuse tooling.

Create a stack frame for them when CONFIG_FRAME_POINTER is enabled and
give them proper ELF function annotations.

Signed-off-by: Josh Poimboeuf 
Reviewed-by: Borislav Petkov 
Acked-by: Pavel Machek 
Acked-by: Rafael J. Wysocki 
Cc: Andrew Morton 
Cc: Andy Lutomirski 
Cc: Andy Lutomirski 
Cc: Arnaldo Carvalho de Melo 
Cc: Bernd Petrovitsch 
Cc: Borislav Petkov 
Cc: Brian Gerst 
Cc: Chris J Arges 
Cc: Denys Vlasenko 
Cc: H. Peter Anvin 
Cc: Jiri Slaby 
Cc: Linus Torvalds 
Cc: Michal Marek 
Cc: Namhyung Kim 
Cc: Pedro Alves 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: live-patching@vger.kernel.org
Link: http://lkml.kernel.org/r/bdad00205897dc707aebe9e9e39757085e2bf999.1453405861.git.jpoimboe@redhat.com
Signed-off-by: Ingo Molnar

x86/pm: Introduce quirk framework to save/restore extra MSR registers around suspend/resume

2015-11-26T09:04:53+00:00

A bug was reported that on certain Broadwell platforms, after
resuming from S3, the CPU is running at an anomalously low
speed.

It turns out that the BIOS has modified the value of the
THERM_CONTROL register during S3, and changed it from 0 to 0x10,
thus enabled clock modulation(bit4), but with undefined CPU Duty
Cycle(bit1:3) - which causes the problem.

Here is a simple scenario to reproduce the issue:

 1. Boot up the system
 2. Get MSR 0x19a, it should be 0
 3. Put the system into sleep, then wake it up
 4. Get MSR 0x19a, it shows 0x10, while it should be 0

Although some BIOSen want to change the CPU Duty Cycle during
S3, in our case we don't want the BIOS to do any modification.

Fix this issue by introducing a more generic x86 framework to
save/restore specified MSR registers(THERM_CONTROL in this case)
for suspend/resume. This allows us to fix similar bugs in a much
simpler way in the future.

When the kernel wants to protect certain MSRs during suspending,
we simply add a quirk entry in msr_save_dmi_table, and customize
the MSR registers inside the quirk callback, for example:

  u32 msr_id_need_to_save[] = {MSR_ID0, MSR_ID1, MSR_ID2...};

and the quirk mechanism ensures that, once resumed from suspend,
the MSRs indicated by these IDs will be restored to their
original, pre-suspend values.

Since both 64-bit and 32-bit kernels are affected, this patch
covers the common 64/32-bit suspend/resume code path. And
because the MSRs specified by the user might not be available or
readable in any situation, we use rdmsrl_safe() to safely save
these MSRs.

Reported-and-tested-by: Marcin Kaszewski 
Signed-off-by: Chen Yu 
Acked-by: Rafael J. Wysocki 
Acked-by: Pavel Machek 
Cc: Andy Lutomirski 
Cc: Borislav Petkov 
Cc: Brian Gerst 
Cc: Denys Vlasenko 
Cc: H. Peter Anvin 
Cc: Linus Torvalds 
Cc: Peter Zijlstra 
Cc: Thomas Gleixner 
Cc: bp@suse.de
Cc: len.brown@intel.com
Cc: linux@horizon.com
Cc: luto@kernel.org
Cc: rjw@rjwysocki.net
Link: http://lkml.kernel.org/r/c9abdcbc173dd2f57e8990e304376f19287e92ba.1448382971.git.yu.c.chen@intel.com
[ More edits to the naming of data structures. ]
Signed-off-by: Ingo Molnar