Currently fpu__activate_fpstate() is used for two distinct purposes:
- read access by ptrace and core dumping, where in the core dumping
case the current task's FPU state may be examined as well.
- write access by ptrace, which modifies FPU registers and expects
the modified registers to be reloaded on the next context switch.
Split out the reading side into fpu__activate_fpstate_read().
( Note that this is just a pure duplication of fpu__activate_fpstate()
for the time being, we'll optimize the new function in the next patch. )
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Bobby Powers <bobbypowers@gmail.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Bobby Powers reported the following FPU warning during ELF coredumping:
WARNING: CPU: 0 PID: 27452 at arch/x86/kernel/fpu/core.c:324 fpu__activate_stopped+0x8a/0xa0()
This warning unearthed an invalid assumption about fpu__activate_stopped()
that I added in:
67e97fc2ec ("x86/fpu: Rename init_fpu() to fpu__unlazy_stopped() and add debugging check")
the old init_fpu() function had an (intentional but obscure) side effect:
when FPU registers are accessed for the current task, for reading, then
it synchronized live in-register FPU state with the fpstate by saving it.
So fix this bug by saving the FPU if we are the current task. We'll
still warn in fpu__save() if this is called for not yet stopped
child tasks, so the debugging check is still preserved.
Also rename the function to fpu__activate_fpstate(), because it's not
exclusively used for stopped tasks, but for the current task as well.
( Note that this bug calls for a cleaner separation of access-for-read
and access-for-modification FPU methods, but we'll do that in separate
patches. )
Reported-by: Bobby Powers <bobbypowers@gmail.com>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The copy_xregs_to_kernel*() and copy_kernel_to_xregs*() functions are used
to copy FPU registers to kernel memory and vice versa.
They are never expected to fail, yet they have a return code, mostly because
that way they can share the assembly macros with the copy*user*() functions.
This error code is then silently ignored by the context switching
and other code - which made the bug in:
b8c1b8ea7b ("x86/fpu: Fix FPU state save area alignment bug")
harder to fix than necessary.
So remove the return values and check for no faults when FPU debugging
is enabled in the .config.
This improves the eagerfpu context switching fast path by a couple of
instructions, when FPU debugging is disabled:
ffffffff810407fa: 89 c2 mov %eax,%edx
ffffffff810407fc: 48 0f ae 2f xrstor64 (%rdi)
ffffffff81040800: 31 c0 xor %eax,%eax
-ffffffff81040802: eb 0a jmp ffffffff8104080e <__switch_to+0x321>
+ffffffff81040802: eb 16 jmp ffffffff8104081a <__switch_to+0x32d>
ffffffff81040804: 31 c0 xor %eax,%eax
ffffffff81040806: 48 0f ae 8b c0 05 00 fxrstor64 0x5c0(%rbx)
ffffffff8104080d: 00
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There's a confusing aspect of how xstate_fault() constraints are
handled by the FPU register/memory copying functions in
fpu/internal.h: they use "0" (0) to signal that the asm code
will not always set 'err' to a valid value.
But 'err' is already initialized to 0 in C code, which is duplicated
by the asm() constraint. Should the initialization value ever be
changed, it might become subtly inconsistent with the not too clear
asm() constraint.
Use 'err' as the value of the input variable instead, to clarify
this all.
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There are two problems with xstate_fault handling:
- The xstate_fault() macro takes an argument, but that's
propagated into the assembly named label as well. This
is technically correct currently but might result in
failures if anytime a more complex argument is used.
So use a separate '_err' name instead for the label.
- All the xstate_fault() using functions have an error
variable named 'err', which is an output variable to
the asm() they are using. The problem is, it's not always
set by the asm(), in which case the compiler might
optimize out its initialization, so that the C variable
'err' might become corrupted after the asm() - confusing
anyone who tries to take advantage of this variable
after the asm(). Mark it an input variable as well.
This is a latent bug currently, but an upcoming debug
patch will make use of 'err'.
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: linux-kernel@vger.kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
On most configs task-struct is cache line aligned, which makes
the XSAVE area's 64-byte required alignment work out fine.
But on some .config's task_struct is aligned only to 16 bytes
(enforced by ARCH_MIN_TASKALIGN), which makes things like
fpu__copy() (that XSAVEOPT uses) not work so well.
I broke this in:
7366ed771f ("x86/fpu: Simplify FPU handling by embedding the fpstate in task_struct (again)")
which embedded the fpstate in the task_struct.
The alignment requirements of the FPU code were originally present
in ARCH_MIN_TASKALIGN, which still has a value of 16, which was the
alignment requirement of the FPU state area prior XSAVE. But this
link was not documented (and not required) and the link got lost
when the FPU state area was made dynamic years ago.
With XSAVEOPT the minimum alignment requirment went up to 64 bytes,
and the embedding of the FPU state area in task_struct exposed it
again - and '16' was not increased to '64'.
So fix this bug, but also try to address the underlying lost link
of information that made it easier to happen:
- document ARCH_MIN_TASKALIGN a bit better
- use alignof() to recover the current alignment requirements.
This would work in the future as well, should the alignment
requirements go up to 128 bytes with things like AVX512.
( We should probably also use the vSMP alignment rules for all
of x86, but that's for another patch. )
Reported-by: Peter Zijlstra <peterz@infradead.org>
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The early_idt_handlers asm code generates an array of entry
points spaced nine bytes apart. It's not really clear from that
code or from the places that reference it what's going on, and
the code only works in the first place because GAS never
generates two-byte JMP instructions when jumping to global
labels.
Clean up the code to generate the correct array stride (member size)
explicitly. This should be considerably more robust against
screw-ups, as GAS will warn if a .fill directive has a negative
count. Using '. =' to advance would have been even more robust
(it would generate an actual error if it tried to move
backwards), but it would pad with nulls, confusing anyone who
tries to disassemble the code. The new scheme should be much
clearer to future readers.
While we're at it, improve the comments and rename the array and
common code.
Binutils may start relaxing jumps to non-weak labels. If so,
this change will fix our build, and we may need to backport this
change.
Before, on x86_64:
0000000000000000 <early_idt_handlers>:
0: 6a 00 pushq $0x0
2: 6a 00 pushq $0x0
4: e9 00 00 00 00 jmpq 9 <early_idt_handlers+0x9>
5: R_X86_64_PC32 early_idt_handler-0x4
...
48: 66 90 xchg %ax,%ax
4a: 6a 08 pushq $0x8
4c: e9 00 00 00 00 jmpq 51 <early_idt_handlers+0x51>
4d: R_X86_64_PC32 early_idt_handler-0x4
...
117: 6a 00 pushq $0x0
119: 6a 1f pushq $0x1f
11b: e9 00 00 00 00 jmpq 120 <early_idt_handler>
11c: R_X86_64_PC32 early_idt_handler-0x4
After:
0000000000000000 <early_idt_handler_array>:
0: 6a 00 pushq $0x0
2: 6a 00 pushq $0x0
4: e9 14 01 00 00 jmpq 11d <early_idt_handler_common>
...
48: 6a 08 pushq $0x8
4a: e9 d1 00 00 00 jmpq 120 <early_idt_handler_common>
4f: cc int3
50: cc int3
...
117: 6a 00 pushq $0x0
119: 6a 1f pushq $0x1f
11b: eb 03 jmp 120 <early_idt_handler_common>
11d: cc int3
11e: cc int3
11f: cc int3
Signed-off-by: Andy Lutomirski <luto@kernel.org>
Acked-by: H. Peter Anvin <hpa@linux.intel.com>
Cc: Binutils <binutils@sourceware.org>
Cc: Borislav Petkov <bp@alien8.de>
Cc: H.J. Lu <hjl.tools@gmail.com>
Cc: Jan Beulich <JBeulich@suse.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Link: http://lkml.kernel.org/r/ac027962af343b0c599cbfcf50b945ad2ef3d7a8.1432336324.git.luto@kernel.org
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Pull KVM fixes from Paolo Bonzini:
"This includes a fix for two oopses, one on PPC and on x86.
The rest is fixes for bugs with newer Intel processors"
* tag 'for-linus' of git://git.kernel.org/pub/scm/virt/kvm/kvm:
kvm/fpu: Enable eager restore kvm FPU for MPX
Revert "KVM: x86: drop fpu_activate hook"
kvm: fix crash in kvm_vcpu_reload_apic_access_page
KVM: MMU: fix SMAP virtualization
KVM: MMU: fix CR4.SMEP=1, CR0.WP=0 with shadow pages
KVM: MMU: fix smap permission check
KVM: PPC: Book3S HV: Fix list traversal in error case
The MPX feature requires eager KVM FPU restore support. We have verified
that MPX cannot work correctly with the current lazy KVM FPU restore
mechanism. Eager KVM FPU restore should be enabled if the MPX feature is
exposed to VM.
Signed-off-by: Yang Zhang <yang.z.zhang@intel.com>
Signed-off-by: Liang Li <liang.z.li@intel.com>
[Also activate the FPU on AMD processors. - Paolo]
Cc: stable@vger.kernel.org
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
KVM may turn a user page to a kernel page when kernel writes a readonly
user page if CR0.WP = 1. This shadow page entry will be reused after
SMAP is enabled so that kernel is allowed to access this user page
Fix it by setting SMAP && !CR0.WP into shadow page's role and reset mmu
once CR4.SMAP is updated
Signed-off-by: Xiao Guangrong <guangrong.xiao@linux.intel.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
fpu/internal.h has grown organically, with not much high level structure,
which hurts its readability.
Organize the various definitions into 5 sections:
- high level FPU state functions
- FPU/CPU feature flag helpers
- fpstate handling functions
- FPU context switching helpers
- misc helper functions
Other related changes:
- Move MXCSR_DEFAULT to fpu/types.h.
- drop the unused X87_FSW_ES define
No change in functionality.
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
There are various internal FPU state debugging checks that never
trigger in practice, but which are useful for FPU code development.
Separate these out into CONFIG_X86_DEBUG_FPU=y, and also add a
couple of new ones.
The size difference is about 0.5K of code on defconfig:
text data bss filename
15028906 2578816 1638400 vmlinux
15029430 2578816 1638400 vmlinux
( Keep this enabled by default until the new FPU code is debugged. )
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
Improve the memory layout of 'struct fpu':
- change ->fpregs_active from 'int' to 'char' - it's just a single flag
and modern x86 CPUs can do efficient byte accesses.
- pack related fields closer to each other: often 'fpu->state' will not be
touched, while the other fields will - so pack them into a group.
Also add comments to each field, describing their purpose, and add
some background information about lazy restores.
Also fix an obsolete, lazy switching related comment in fpu_copy()'s description.
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
So the handling of init_xstate_ctx has a layering violation: both
'struct xsave_struct' and 'union thread_xstate' have a
'struct i387_fxsave_struct' member:
xsave_struct::i387
thread_xstate::fxsave
The handling of init_xstate_ctx is generic, it is used on all
CPUs, with or without XSAVE instruction. So it's confusing how
the generic code passes around and handles an XSAVE specific
format.
What we really want is for init_xstate_ctx to be a proper
fpstate and we use its ::fxsave and ::xsave members, as
appropriate.
Since the xsave_struct::i387 and thread_xstate::fxsave aliases
each other this is not a functional problem.
So implement this, and move init_xstate_ctx to the generic FPU
code in the process.
Also, since init_xstate_ctx is not XSAVE specific anymore,
rename it to init_fpstate, and mark it __read_mostly,
because it's only modified once during bootup, and used
as a reference fpstate later on.
There's no change in functionality.
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
With recent cleanups and fixes the fpu__reset() and fpu__clear()
functions have become almost identical in functionality: the only
difference is that fpu__reset() assumed that the fpstate
was already active in the eagerfpu case, while fpu__clear()
activated it if it was inactive.
This distinction almost never matters, the only case where such
fpstate activation happens if if the init thread (PID 1) gets exec()-ed
for the first time.
So keep fpu__clear() and change all fpu__reset() uses to
fpu__clear() to simpify the logic.
( In a later patch we'll further simplify fpu__clear() by making
sure that all contexts it is called on are already active. )
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
The fpu__*() methods are closely related, but they are defined
in scattered places within the FPU code.
Concentrate them, and also uninline fpu__save(), fpu__drop()
and fpu__reset() to save about 5K of kernel text on 64-bit kernels:
text data bss dec filename
14113063 2575280 1634304 18322647 vmlinux.before
14108070 2575280 1634304 18317654 vmlinux.after
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
drop_fpu() and fpu_reset_state() are similar in functionality
and in scope, yet this is not apparent from their names.
drop_fpu() deactivates FPU contents (both the fpregs and the fpstate),
but leaves register contents intact in the eager-FPU case, mostly as an
optimization. It disables fpregs in the lazy FPU case. The drop_fpu()
method can be used to destroy FPU state in an optimized way, when we
know that a new state will be loaded before user-space might see
any remains of the old FPU state:
- such as in sys_exit()'s exit_thread() where we know this task
won't execute any user-space instructions anymore and the
next context switch cleans up the FPU. The old FPU state
might still be around in the eagerfpu case but won't be
saved.
- in __restore_xstate_sig(), where we use drop_fpu() before
copying a new state into the fpstate and activating that one.
No user-pace instructions can execute between those steps.
- in sys_execve()'s fpu__clear(): there we use drop_fpu() in
the !eagerfpu case, where it's equivalent to a full reinit.
fpu_reset_state() is a stronger version of drop_fpu(): both in
the eagerfpu and the lazy-FPU case it guarantees that fpregs
are reinitialized to init state. This method is used in cases
where we need a full reset:
- handle_signal() uses fpu_reset_state() to reset the FPU state
to init before executing a user-space signal handler. While we
have already saved the original FPU state at this point, and
always restore the original state, the signal handling code
still has to do this reinit, because signals may interrupt
any user-space instruction, and the FPU might be in various
intermediate states (such as an unbalanced x87 stack) that is
not immediately usable for general C signal handler code.
- __restore_xstate_sig() uses fpu_reset_state() when the signal
frame has no FP context. Since the signal handler may have
modified the FPU state, it gets reset back to init state.
- in another branch __restore_xstate_sig() uses fpu_reset_state()
to handle a restoration error: when restore_user_xstate() fails
to restore FPU state and we might have inconsistent FPU data,
fpu_reset_state() is used to reset it back to a known good
state.
- __kernel_fpu_end() uses fpu_reset_state() in an error branch.
This is in a 'must not trigger' error branch, so on bug-free
kernels this never triggers.
- fpu__restore() uses fpu_reset_state() in an error path
as well: if the fpstate was set up with invalid FPU state
(via ptrace or via a signal handler), then it's reset back
to init state.
- likewise, the scheduler's switch_fpu_finish() uses it in a
restoration error path too.
Move both drop_fpu() and fpu_reset_state() to the fpu__*() namespace
and harmonize their naming with their function:
fpu__drop()
fpu__reset()
This clearly shows that both methods operate on the full state of the
FPU, just like fpu__restore().
Also add comments to explain what each function does.
Cc: Andy Lutomirski <luto@amacapital.net>
Cc: Borislav Petkov <bp@alien8.de>
Cc: Dave Hansen <dave.hansen@linux.intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: H. Peter Anvin <hpa@zytor.com>
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Ingo Molnar <mingo@kernel.org>
