linux/arch/arm/vfp/vfphw.S
Russell King f8f2a8522a ARM: vfp: fix a hole in VFP thread migration
Fix a hole in the VFP thread migration.  Lets define two threads.

Thread 1, we'll call 'interesting_thread' which is a thread which is
running on CPU0, using VFP (so vfp_current_hw_state[0] =
&interesting_thread->vfpstate) and gets migrated off to CPU1, where
it continues execution of VFP instructions.

Thread 2, we'll call 'new_cpu0_thread' which is the thread which takes
over on CPU0.  This has also been using VFP, and last used VFP on CPU0,
but doesn't use it again.

The following code will be executed twice:

		cpu = thread->cpu;

		/*
		 * On SMP, if VFP is enabled, save the old state in
		 * case the thread migrates to a different CPU. The
		 * restoring is done lazily.
		 */
		if ((fpexc & FPEXC_EN) && vfp_current_hw_state[cpu]) {
			vfp_save_state(vfp_current_hw_state[cpu], fpexc);
			vfp_current_hw_state[cpu]->hard.cpu = cpu;
		}
		/*
		 * Thread migration, just force the reloading of the
		 * state on the new CPU in case the VFP registers
		 * contain stale data.
		 */
		if (thread->vfpstate.hard.cpu != cpu)
			vfp_current_hw_state[cpu] = NULL;

The first execution will be on CPU0 to switch away from 'interesting_thread'.
interesting_thread->cpu will be 0.

So, vfp_current_hw_state[0] points at interesting_thread->vfpstate.
The hardware state will be saved, along with the CPU number (0) that
it was executing on.

'thread' will be 'new_cpu0_thread' with new_cpu0_thread->cpu = 0.
Also, because it was executing on CPU0, new_cpu0_thread->vfpstate.hard.cpu = 0,
and so the thread migration check is not triggered.

This means that vfp_current_hw_state[0] remains pointing at interesting_thread.

The second execution will be on CPU1 to switch _to_ 'interesting_thread'.
So, 'thread' will be 'interesting_thread' and interesting_thread->cpu now
will be 1.  The previous thread executing on CPU1 is not relevant to this
so we shall ignore that.

We get to the thread migration check.  Here, we discover that
interesting_thread->vfpstate.hard.cpu = 0, yet interesting_thread->cpu is
now 1, indicating thread migration.  We set vfp_current_hw_state[1] to
NULL.

So, at this point vfp_current_hw_state[] contains the following:

[0] = &interesting_thread->vfpstate
[1] = NULL

Our interesting thread now executes a VFP instruction, takes a fault
which loads the state into the VFP hardware.  Now, through the assembly
we now have:

[0] = &interesting_thread->vfpstate
[1] = &interesting_thread->vfpstate

CPU1 stops due to ptrace (and so saves its VFP state) using the thread
switch code above), and CPU0 calls vfp_sync_hwstate().

	if (vfp_current_hw_state[cpu] == &thread->vfpstate) {
		vfp_save_state(&thread->vfpstate, fpexc | FPEXC_EN);

BANG, we corrupt interesting_thread's VFP state by overwriting the
more up-to-date state saved by CPU1 with the old VFP state from CPU0.

Fix this by ensuring that we have sane semantics for the various state
describing variables:

1. vfp_current_hw_state[] points to the current owner of the context
   information stored in each CPUs hardware, or NULL if that state
   information is invalid.
2. thread->vfpstate.hard.cpu always contains the most recent CPU number
   which the state was loaded into or NR_CPUS if no CPU owns the state.

So, for a particular CPU to be a valid owner of the VFP state for a
particular thread t, two things must be true:

 vfp_current_hw_state[cpu] == &t->vfpstate && t->vfpstate.hard.cpu == cpu.

and that is valid from the moment a CPU loads the saved VFP context
into the hardware.  This gives clear and consistent semantics to
interpreting these variables.

This patch also fixes thread copying, ensuring that t->vfpstate.hard.cpu
is invalidated, otherwise CPU0 may believe it was the last owner.  The
hole can happen thus:

- thread1 runs on CPU2 using VFP, migrates to CPU3, exits and thread_info
  freed.
- New thread allocated from a previously running thread on CPU2, reusing
  memory for thread1 and copying vfp.hard.cpu.

At this point, the following are true:

	new_thread1->vfpstate.hard.cpu == 2
	&new_thread1->vfpstate == vfp_current_hw_state[2]

Lastly, this also addresses thread flushing in a similar way to thread
copying.  Hole is:

- thread runs on CPU0, using VFP, migrates to CPU1 but does not use VFP.
- thread calls execve(), so thread flush happens, leaving
  vfp_current_hw_state[0] intact.  This vfpstate is memset to 0 causing
  thread->vfpstate.hard.cpu = 0.
- thread migrates back to CPU0 before using VFP.

At this point, the following are true:

	thread->vfpstate.hard.cpu == 0
	&thread->vfpstate == vfp_current_hw_state[0]

Signed-off-by: Russell King <rmk+kernel@arm.linux.org.uk>
2011-07-09 17:22:12 +01:00

310 lines
8.2 KiB
ArmAsm

/*
* linux/arch/arm/vfp/vfphw.S
*
* Copyright (C) 2004 ARM Limited.
* Written by Deep Blue Solutions Limited.
*
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License version 2 as
* published by the Free Software Foundation.
*
* This code is called from the kernel's undefined instruction trap.
* r9 holds the return address for successful handling.
* lr holds the return address for unrecognised instructions.
* r10 points at the start of the private FP workspace in the thread structure
* sp points to a struct pt_regs (as defined in include/asm/proc/ptrace.h)
*/
#include <asm/thread_info.h>
#include <asm/vfpmacros.h>
#include "../kernel/entry-header.S"
.macro DBGSTR, str
#ifdef DEBUG
stmfd sp!, {r0-r3, ip, lr}
add r0, pc, #4
bl printk
b 1f
.asciz "<7>VFP: \str\n"
.balign 4
1: ldmfd sp!, {r0-r3, ip, lr}
#endif
.endm
.macro DBGSTR1, str, arg
#ifdef DEBUG
stmfd sp!, {r0-r3, ip, lr}
mov r1, \arg
add r0, pc, #4
bl printk
b 1f
.asciz "<7>VFP: \str\n"
.balign 4
1: ldmfd sp!, {r0-r3, ip, lr}
#endif
.endm
.macro DBGSTR3, str, arg1, arg2, arg3
#ifdef DEBUG
stmfd sp!, {r0-r3, ip, lr}
mov r3, \arg3
mov r2, \arg2
mov r1, \arg1
add r0, pc, #4
bl printk
b 1f
.asciz "<7>VFP: \str\n"
.balign 4
1: ldmfd sp!, {r0-r3, ip, lr}
#endif
.endm
@ VFP hardware support entry point.
@
@ r0 = faulted instruction
@ r2 = faulted PC+4
@ r9 = successful return
@ r10 = vfp_state union
@ r11 = CPU number
@ lr = failure return
ENTRY(vfp_support_entry)
DBGSTR3 "instr %08x pc %08x state %p", r0, r2, r10
VFPFMRX r1, FPEXC @ Is the VFP enabled?
DBGSTR1 "fpexc %08x", r1
tst r1, #FPEXC_EN
bne look_for_VFP_exceptions @ VFP is already enabled
DBGSTR1 "enable %x", r10
ldr r3, vfp_current_hw_state_address
orr r1, r1, #FPEXC_EN @ user FPEXC has the enable bit set
ldr r4, [r3, r11, lsl #2] @ vfp_current_hw_state pointer
bic r5, r1, #FPEXC_EX @ make sure exceptions are disabled
cmp r4, r10 @ this thread owns the hw context?
#ifndef CONFIG_SMP
@ For UP, checking that this thread owns the hw context is
@ sufficient to determine that the hardware state is valid.
beq vfp_hw_state_valid
@ On UP, we lazily save the VFP context. As a different
@ thread wants ownership of the VFP hardware, save the old
@ state if there was a previous (valid) owner.
VFPFMXR FPEXC, r5 @ enable VFP, disable any pending
@ exceptions, so we can get at the
@ rest of it
DBGSTR1 "save old state %p", r4
cmp r4, #0 @ if the vfp_current_hw_state is NULL
beq vfp_reload_hw @ then the hw state needs reloading
VFPFSTMIA r4, r5 @ save the working registers
VFPFMRX r5, FPSCR @ current status
#ifndef CONFIG_CPU_FEROCEON
tst r1, #FPEXC_EX @ is there additional state to save?
beq 1f
VFPFMRX r6, FPINST @ FPINST (only if FPEXC.EX is set)
tst r1, #FPEXC_FP2V @ is there an FPINST2 to read?
beq 1f
VFPFMRX r8, FPINST2 @ FPINST2 if needed (and present)
1:
#endif
stmia r4, {r1, r5, r6, r8} @ save FPEXC, FPSCR, FPINST, FPINST2
vfp_reload_hw:
#else
@ For SMP, if this thread does not own the hw context, then we
@ need to reload it. No need to save the old state as on SMP,
@ we always save the state when we switch away from a thread.
bne vfp_reload_hw
@ This thread has ownership of the current hardware context.
@ However, it may have been migrated to another CPU, in which
@ case the saved state is newer than the hardware context.
@ Check this by looking at the CPU number which the state was
@ last loaded onto.
ldr ip, [r10, #VFP_CPU]
teq ip, r11
beq vfp_hw_state_valid
vfp_reload_hw:
@ We're loading this threads state into the VFP hardware. Update
@ the CPU number which contains the most up to date VFP context.
str r11, [r10, #VFP_CPU]
VFPFMXR FPEXC, r5 @ enable VFP, disable any pending
@ exceptions, so we can get at the
@ rest of it
#endif
DBGSTR1 "load state %p", r10
str r10, [r3, r11, lsl #2] @ update the vfp_current_hw_state pointer
@ Load the saved state back into the VFP
VFPFLDMIA r10, r5 @ reload the working registers while
@ FPEXC is in a safe state
ldmia r10, {r1, r5, r6, r8} @ load FPEXC, FPSCR, FPINST, FPINST2
#ifndef CONFIG_CPU_FEROCEON
tst r1, #FPEXC_EX @ is there additional state to restore?
beq 1f
VFPFMXR FPINST, r6 @ restore FPINST (only if FPEXC.EX is set)
tst r1, #FPEXC_FP2V @ is there an FPINST2 to write?
beq 1f
VFPFMXR FPINST2, r8 @ FPINST2 if needed (and present)
1:
#endif
VFPFMXR FPSCR, r5 @ restore status
@ The context stored in the VFP hardware is up to date with this thread
vfp_hw_state_valid:
tst r1, #FPEXC_EX
bne process_exception @ might as well handle the pending
@ exception before retrying branch
@ out before setting an FPEXC that
@ stops us reading stuff
VFPFMXR FPEXC, r1 @ restore FPEXC last
sub r2, r2, #4
str r2, [sp, #S_PC] @ retry the instruction
#ifdef CONFIG_PREEMPT
get_thread_info r10
ldr r4, [r10, #TI_PREEMPT] @ get preempt count
sub r11, r4, #1 @ decrement it
str r11, [r10, #TI_PREEMPT]
#endif
mov pc, r9 @ we think we have handled things
look_for_VFP_exceptions:
@ Check for synchronous or asynchronous exception
tst r1, #FPEXC_EX | FPEXC_DEX
bne process_exception
@ On some implementations of the VFP subarch 1, setting FPSCR.IXE
@ causes all the CDP instructions to be bounced synchronously without
@ setting the FPEXC.EX bit
VFPFMRX r5, FPSCR
tst r5, #FPSCR_IXE
bne process_exception
@ Fall into hand on to next handler - appropriate coproc instr
@ not recognised by VFP
DBGSTR "not VFP"
#ifdef CONFIG_PREEMPT
get_thread_info r10
ldr r4, [r10, #TI_PREEMPT] @ get preempt count
sub r11, r4, #1 @ decrement it
str r11, [r10, #TI_PREEMPT]
#endif
mov pc, lr
process_exception:
DBGSTR "bounce"
mov r2, sp @ nothing stacked - regdump is at TOS
mov lr, r9 @ setup for a return to the user code.
@ Now call the C code to package up the bounce to the support code
@ r0 holds the trigger instruction
@ r1 holds the FPEXC value
@ r2 pointer to register dump
b VFP_bounce @ we have handled this - the support
@ code will raise an exception if
@ required. If not, the user code will
@ retry the faulted instruction
ENDPROC(vfp_support_entry)
ENTRY(vfp_save_state)
@ Save the current VFP state
@ r0 - save location
@ r1 - FPEXC
DBGSTR1 "save VFP state %p", r0
VFPFSTMIA r0, r2 @ save the working registers
VFPFMRX r2, FPSCR @ current status
tst r1, #FPEXC_EX @ is there additional state to save?
beq 1f
VFPFMRX r3, FPINST @ FPINST (only if FPEXC.EX is set)
tst r1, #FPEXC_FP2V @ is there an FPINST2 to read?
beq 1f
VFPFMRX r12, FPINST2 @ FPINST2 if needed (and present)
1:
stmia r0, {r1, r2, r3, r12} @ save FPEXC, FPSCR, FPINST, FPINST2
mov pc, lr
ENDPROC(vfp_save_state)
.align
vfp_current_hw_state_address:
.word vfp_current_hw_state
.macro tbl_branch, base, tmp, shift
#ifdef CONFIG_THUMB2_KERNEL
adr \tmp, 1f
add \tmp, \tmp, \base, lsl \shift
mov pc, \tmp
#else
add pc, pc, \base, lsl \shift
mov r0, r0
#endif
1:
.endm
ENTRY(vfp_get_float)
tbl_branch r0, r3, #3
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: mrc p10, 0, r0, c\dr, c0, 0 @ fmrs r0, s0
mov pc, lr
.org 1b + 8
1: mrc p10, 0, r0, c\dr, c0, 4 @ fmrs r0, s1
mov pc, lr
.org 1b + 8
.endr
ENDPROC(vfp_get_float)
ENTRY(vfp_put_float)
tbl_branch r1, r3, #3
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: mcr p10, 0, r0, c\dr, c0, 0 @ fmsr r0, s0
mov pc, lr
.org 1b + 8
1: mcr p10, 0, r0, c\dr, c0, 4 @ fmsr r0, s1
mov pc, lr
.org 1b + 8
.endr
ENDPROC(vfp_put_float)
ENTRY(vfp_get_double)
tbl_branch r0, r3, #3
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: fmrrd r0, r1, d\dr
mov pc, lr
.org 1b + 8
.endr
#ifdef CONFIG_VFPv3
@ d16 - d31 registers
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: mrrc p11, 3, r0, r1, c\dr @ fmrrd r0, r1, d\dr
mov pc, lr
.org 1b + 8
.endr
#endif
@ virtual register 16 (or 32 if VFPv3) for compare with zero
mov r0, #0
mov r1, #0
mov pc, lr
ENDPROC(vfp_get_double)
ENTRY(vfp_put_double)
tbl_branch r2, r3, #3
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: fmdrr d\dr, r0, r1
mov pc, lr
.org 1b + 8
.endr
#ifdef CONFIG_VFPv3
@ d16 - d31 registers
.irp dr,0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15
1: mcrr p11, 3, r0, r1, c\dr @ fmdrr r0, r1, d\dr
mov pc, lr
.org 1b + 8
.endr
#endif
ENDPROC(vfp_put_double)