Re: [RFC PATCH] kprobes: arm: enable OPTPROBES for arm 32

From: Wang Nan
Date: Wed Aug 06 2014 - 02:25:41 EST


Thank you for your comments. I'm waiting for your test result and preparing
the next version. Some response below.

On 2014/8/6 12:44, Masami Hiramatsu wrote:
> (2014/08/05 16:28), Wang Nan wrote:
>> This patch introduce kprobeopt for ARM 32.
>
> Thanks you for the great work! This looks fine for me. I'd like to
> test it on my cortex-a9 board.
>
>>
>> Limitations:
>> - Currently only kernel compiled with ARM ISA is supported.
>
> It is good to start from ARM ISA, thumb2 is more complex.
>
>> - offset between probe point and kprobe pre_handler must not larger

My mistake: not "offset between probe point and kprobe pre_handler". It
should be "offset between probe point and optinsn slot".

64MiB is still enough. optinsn slot is allocated using module_alloc. On arm,
its address will reside in MODULES_VADDR - MODULES_END.

>> than 64MiB. Masami Hiramatsu suggests replacing 2 words, it will
>> make things complex. Futher patch can make such optimization.
>
> OK, it seems that arm32 has closed memory areas for modules and
> kernel text. So 64MiB is enough to cover both.
>
> modules : 0x7f000000 - 0x80000000 ( 16 MB)
> .text : 0x80008000 - 0x8070138c (7141 kB)
>
>> I have tested this patch on qemu vexpress a9 platform.
>>
>> Kprobe opt on ARM is relatively simpler than kprobe opt on x86 because
>> ARM instruction is always 4 bytes aligned. This patch replace probed
^^^^^
and 4 bytes long.

>> instruction by a 'b', branch to trampoline code and then calls
>> optimized_callback(). optimized_callback() calls kprobe_handler() to
>> execute kprobe handler. It also emulate/simulate replaced instruction.
>
> I see, this simplicity removes all basic-block analysis from the code,
> because it always replaces just one instruction.
>
>> When unregistering kprobe, the deferred manner of unoptimizer may leave
>> branch instruction before optimizer is called. Different from x86_64,
>> which only copy the probed insn after optprobe_template_end and
>> reexecute them, this patch call singlestep to emulate/simulate the insn
>> directly. Futher patch can optimize this behavior.
>>
>> Signed-off-by: Wang Nan <wangnan0@xxxxxxxxxx>
>> Cc: Masami Hiramatsu <masami.hiramatsu.pt@xxxxxxxxxxx>
>> ---
>> arch/arm/Kconfig | 1 +
>> arch/arm/include/asm/kprobes.h | 20 +++
>> arch/arm/kernel/Makefile | 1 +
>> arch/arm/kernel/kprobes-opt.c | 322 +++++++++++++++++++++++++++++++++++++++++
>> 4 files changed, 344 insertions(+)
>> create mode 100644 arch/arm/kernel/kprobes-opt.c
>>
>> diff --git a/arch/arm/Kconfig b/arch/arm/Kconfig
>> index 290f02ee..2106918 100644
>> --- a/arch/arm/Kconfig
>> +++ b/arch/arm/Kconfig
>> @@ -57,6 +57,7 @@ config ARM
>> select HAVE_MEMBLOCK
>> select HAVE_MOD_ARCH_SPECIFIC if ARM_UNWIND
>> select HAVE_OPROFILE if (HAVE_PERF_EVENTS)
>> + select HAVE_OPTPROBES if (!THUMB2_KERNEL)
>> select HAVE_PERF_EVENTS
>> select HAVE_PERF_REGS
>> select HAVE_PERF_USER_STACK_DUMP
>> diff --git a/arch/arm/include/asm/kprobes.h b/arch/arm/include/asm/kprobes.h
>> index 49fa0df..0f4e5c0 100644
>> --- a/arch/arm/include/asm/kprobes.h
>> +++ b/arch/arm/include/asm/kprobes.h
>> @@ -51,5 +51,25 @@ int kprobe_fault_handler(struct pt_regs *regs, unsigned int fsr);
>> int kprobe_exceptions_notify(struct notifier_block *self,
>> unsigned long val, void *data);
>>
>> +/* optinsn template addresses */
>> +extern __visible kprobe_opcode_t optprobe_template_entry;
>> +extern __visible kprobe_opcode_t optprobe_template_val;
>> +extern __visible kprobe_opcode_t optprobe_template_call;
>> +extern __visible kprobe_opcode_t optprobe_template_end;
>> +
>> +#define MAX_OPTIMIZED_LENGTH (4)
>> +#define MAX_OPTINSN_SIZE \
>> + (((unsigned long)&optprobe_template_end - \
>> + (unsigned long)&optprobe_template_entry))
>> +#define RELATIVEJUMP_SIZE (4)
>> +
>> +struct arch_optimized_insn {
>> + /* copy of the original instructions */
>> + kprobe_opcode_t copied_insn[RELATIVEJUMP_SIZE];
>> + /* detour code buffer */
>> + kprobe_opcode_t *insn;
>> + /* the size of instructions copied to detour code buffer */
>> + size_t size;
>
> you don't need arch_optimized_insn.size. just add a comment like below.
> /* we always copies one instruction on arm32, size always be 4 */
>
>> +};
>>
>> #endif /* _ARM_KPROBES_H */
>> diff --git a/arch/arm/kernel/Makefile b/arch/arm/kernel/Makefile
>> index 38ddd9f..918ee34 100644
>> --- a/arch/arm/kernel/Makefile
>> +++ b/arch/arm/kernel/Makefile
>> @@ -57,6 +57,7 @@ ifdef CONFIG_THUMB2_KERNEL
>> obj-$(CONFIG_KPROBES) += kprobes-thumb.o probes-thumb.o
>> else
>> obj-$(CONFIG_KPROBES) += kprobes-arm.o probes-arm.o
>> +obj-$(CONFIG_OPTPROBES) += kprobes-opt.o
>> endif
>> obj-$(CONFIG_ARM_KPROBES_TEST) += test-kprobes.o
>> test-kprobes-objs := kprobes-test.o
>> diff --git a/arch/arm/kernel/kprobes-opt.c b/arch/arm/kernel/kprobes-opt.c
>> new file mode 100644
>> index 0000000..a4aa7ed
>> --- /dev/null
>> +++ b/arch/arm/kernel/kprobes-opt.c
>> @@ -0,0 +1,322 @@
>> +/*
>> + * Kernel Probes Jump Optimization (Optprobes)
>> + *
>> + * This program is free software; you can redistribute it and/or modify
>> + * it under the terms of the GNU General Public License as published by
>> + * the Free Software Foundation; either version 2 of the License, or
>> + * (at your option) any later version.
>> + *
>> + * This program is distributed in the hope that it will be useful,
>> + * but WITHOUT ANY WARRANTY; without even the implied warranty of
>> + * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
>> + * GNU General Public License for more details.
>> + *
>> + * You should have received a copy of the GNU General Public License
>> + * along with this program; if not, write to the Free Software
>> + * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA.
>> + *
>> + * Copyright (C) IBM Corporation, 2002, 2004
>> + * Copyright (C) Hitachi Ltd., 2012
>> + * Copyright (C) Huawei Inc., 2014
>> + */
>> +
>> +#include <linux/kprobes.h>
>> +#include <linux/jump_label.h>
>> +#include <asm/kprobes.h>
>> +#include <asm/cacheflush.h>
>> +#include "patch.h"
>> +
>> +unsigned long
>> +__recover_optprobed_insn(kprobe_opcode_t *buf, unsigned long addr)
>> +{
>> + struct optimized_kprobe *op;
>> + struct kprobe *kp;
>> + int i;
>> + long offs;
>> + for (i = 0; i < RELATIVEJUMP_SIZE; i++) {
>> + kp = get_kprobe((void *)addr - i);
>> + if (!kp || !kprobe_optimized(kp))
>> + continue;
>> + op = container_of(kp, struct optimized_kprobe, kp);
>> + /* If op->list is not empty, op is under optimizing */
>> + if (list_empty(&op->list))
>> + goto found;
>> + }
>> +
>> + return addr;
>> +
>> +found:
>> + offs = addr - (unsigned long)kp->addr;
>> + memcpy(buf, op->optinsn.copied_insn + offs, RELATIVEJUMP_SIZE - offs);
>> + return (unsigned long)buf;
>> +}
>> +
>> +static unsigned long
>> +__recover_probed_insn(kprobe_opcode_t *buf, unsigned long addr)
>> +{
>> + struct kprobe *kp;
>> +
>> + kp = get_kprobe((void *)addr);
>> + /* There is no probe, return original address */
>> + if (!kp)
>> + return addr;
>> +
>> + /*
>> + * Basically, kp->ainsn.insn has an original instruction.
>> + */
>> + memcpy(buf, kp->ainsn.insn, MAX_INSN_SIZE * sizeof(kprobe_opcode_t));
>> + return (unsigned long)buf;
>> +}
>
> I think those recovering instruction routines should be same code in ARM
> because both of them replace just 1 instruction. No difference are there.
>

This is copied from x86 code. It seems whole recover code can be simplified.

>> +
>> +/*
>> + * Recover the probed instruction at addr for further analysis.
>> + * Caller must lock kprobes by kprobe_mutex, or disable preemption
>> + * for preventing to release referencing kprobes.
>> + */
>> +unsigned long
>> +recover_probed_instruction(kprobe_opcode_t *buf, unsigned long addr)
>> +{
>> + unsigned long __addr;
>> +
>> + __addr = __recover_optprobed_insn(buf, addr);
>> + if (__addr != addr)
>> + return __addr;
>> +
>> + return __recover_probed_insn(buf, addr);
>> +}
>> +
>> +
>> +/*
>> + * recover kprobed instruction and copy to dest
>> + */
>> +int
>> +__copy_instruction(u8 *dest, u8 *src)
>> +{
>> + kprobe_opcode_t buf[MAX_INSN_SIZE];
>> + memcpy(dest, (void *)recover_probed_instruction(buf, (unsigned long)src),
>> + sizeof(kprobe_opcode_t) * MAX_INSN_SIZE);
>> + return (sizeof(kprobe_opcode_t) * MAX_INSN_SIZE);
>> +}

We only need 1 instruction, so I can remove MAX_INSN_SIZE.
Also, this function will be made static.

>> +
>> +
>> +asm (
>> + ".global optprobe_template_entry\n"
>> + "optprobe_template_entry:\n"
>> +#ifndef CONFIG_THUMB
>> + " sub sp, sp, #80\n"
>> + " stmia sp, {r0 - r14} \n"
>> + " add r3, sp, #80\n"
>> + " str r3, [sp, #52]\n"
>> + " mrs r4, cpsr\n"
>> + " str r4, [sp, #64]\n"
>> + " mov r1, sp\n"
>> + " ldr r0, 1f\n"
>> + " ldr r2, 2f\n"
>> + " blx r2\n"
>> + " ldr r1, [sp, #64]\n"
>> + " msr cpsr_fs, r1\n"
>> + " ldmia sp, {r0 - r15}\n"
>> + ".global optprobe_template_val\n"
>> + "optprobe_template_val:\n"
>> + "1: nop\n"
>> + ".global optprobe_template_call\n"
>> + "optprobe_template_call:\n"
>> + "2: nop\n"
>> +#else /* CONFIG_THUMB */
>> +# error optprobe for thumb is not supported.
>
> Can we set CONFIG_THUMB=y without CONFIG_THUMB2_KERNEL ?
>

This ifdef is left here for futher extention. However we can add them back
when we really support thumb. I'll remove it in the next version.

>> +#endif
>> + ".global optprobe_template_end\n"
>> + "optprobe_template_end:\n");
>> +
>> +#define TMPL_VAL_IDX \
>> + ((long)&optprobe_template_val - (long)&optprobe_template_entry)
>> +#define TMPL_CALL_IDX \
>> + ((long)&optprobe_template_call - (long)&optprobe_template_entry)
>> +#define TMPL_END_IDX \
>> + ((long)&optprobe_template_end - (long)&optprobe_template_entry)
>> +
>> +int arch_prepared_optinsn(struct arch_optimized_insn *optinsn)
>> +{
>> + return optinsn->size;
>> +}
>> +
>> +int arch_check_optimized_kprobe(struct optimized_kprobe *op)
>> +{
>> + int i;
>> + struct kprobe *p;
>> +
>> + for (i = 1; i < op->optinsn.size; i++) {
>> + p = get_kprobe(op->kp.addr + i);
>> + if (p && !kprobe_disabled(p))
>> + return -EEXIST;
>> + }
>> +
>> + return 0;
>> +}
>> +
>> +static int can_optimize(unsigned long paddr)
>> +{
>> + unsigned long size = 0, offset = 0;
>> +
>> + /* Lookup symbol including addr */
>> + if (!kallsyms_lookup_size_offset(paddr, &size, &offset))
>> + return 0;
>> +
>> + /* Check there is enough space for a relative jump. */
>> + if (size - offset < RELATIVEJUMP_SIZE)
>> + return 0;
>
> This looks a bit odd. since the offset always be smaller than size,
> or it returns another symbol.
> Since the arm32 optprobe replaces just one instruction, this should
> always return 1.
>

Copied from x86. You are right, for arm kernel we can always optimize an instruction.

>> +
>> + return 1;
>> +}
>> +
>> +/* Free optimized instruction slot */
>> +static
>> +void __arch_remove_optimized_kprobe(struct optimized_kprobe *op, int dirty)
>> +{
>> + if (op->optinsn.insn) {
>> + free_optinsn_slot(op->optinsn.insn, dirty);
>> + op->optinsn.insn = NULL;
>> + op->optinsn.size = 0;
>> + }
>> +}
>> +
>> +extern void kprobe_handler(struct pt_regs *regs);
>> +
>> +static void
>> +optimized_callback(struct optimized_kprobe *op, struct pt_regs *regs)
>> +{
>> + unsigned long flags;
>> +
>> + regs->ARM_pc = (unsigned long)op->kp.addr;
>> + regs->ARM_ORIG_r0 = ~0UL;
>> +
>> +
>> + local_irq_save(flags);
>> + /*
>> + * This is possible if op is under delayed unoptimizing.
>> + * We need simulate the replaced instruction.
>> + */
>> + if (kprobe_disabled(&op->kp)) {
>> + struct kprobe *p = &op->kp;
>> + op->kp.ainsn.insn_singlestep(p->opcode, &p->ainsn, regs);
>> + } else {
>> + kprobe_handler(regs);
>> + }
>
> You don't need brace "{}" for one statement.
> By the way, why don't you call opt_pre_handler()?
>

I use kprobe_handler because it handles instruction emulation.

In addition, I'm not very sure whether skipping the complex checks
in kprobe_handler() is safe or not.

What about:

struct kprobe *p = &op->kp;
opt_pre_handler(p, regs);
op->kp.ainsn.insn_singlestep(p->opcode, &p->ainsn, regs);

?


>> +
>> + local_irq_restore(flags);
>> +}
>> +
>> +int arch_prepare_optimized_kprobe(struct optimized_kprobe *op)
>> +{
>> + u8 *buf;
>> + long rel;
>> + unsigned long val;
>> +
>> + if (!can_optimize((unsigned long)op->kp.addr))
>> + return -EILSEQ;
>> +
>> + op->optinsn.insn = get_optinsn_slot();
>> + if (!op->optinsn.insn)
>> + return -ENOMEM;
>> +
>> + /*
>> + * Verify if the address gap is in 2GB range, because this uses
> ^64MB? :)
>
>> + * a relative jump.
>> + */
>> + rel = (long)op->optinsn.insn - (long)op->kp.addr + RELATIVEJUMP_SIZE;
>> + if ((abs(rel) > 0x3fffffc) || ((abs(rel) & 3) != 0)) {
>> + __arch_remove_optimized_kprobe(op, 0);
>> + return -ERANGE;
>> + }
>> +
>> + buf = (u8 *)op->optinsn.insn;
>> +
>> + op->optinsn.size = RELATIVEJUMP_SIZE;
>> +
>> + /* Copy arch-dep-instance from template */
>> + memcpy(buf, &optprobe_template_entry, TMPL_END_IDX);
>> +
>> + /* Set probe information */
>> + val = (unsigned long)op;
>> + memcpy(buf + TMPL_VAL_IDX, &val, sizeof(val));
>> +
>> + /* Set probe function call */
>> + val = (unsigned long)optimized_callback;
>> + memcpy(buf + TMPL_CALL_IDX, &val, sizeof(val));
>> +
>> + flush_icache_range((unsigned long) buf,
>> + (unsigned long) buf + TMPL_END_IDX +
>> + op->optinsn.size + RELATIVEJUMP_SIZE);
>
> You don't need to flush the area larger than buf + TMPL_END_IDX.
> On x86, Since there is a tail buffer to execute copied instructions
> and a jump, it does that. But arm32 has no such direct-execute tail
> buffer.
>
>
>> + return 0;
>> +}
>> +
>> +static inline int
>> +arm_branch_to_addr(unsigned int *pinst, void *src, void *dest)
>> +{
>> + unsigned int inst = 0xea000000;
>> + long offset = (unsigned long)(dest) -
>> + ((unsigned long)(src) + 8);
>> + if ((offset > 0x3fffffc) || (offset < -0x3fffffc)) {
>> + printk(KERN_WARNING "Failed to instrument %pS to %pS\n", src, dest);
>> + return -EINVAL;
>
> We've already check this above. If you consider something bad, you'd better put BUG() here.
>
> Thank you,
>
>> + }
>> +
>> + inst |= (((unsigned long)offset) >> 2) & (0x00ffffffUL);
>> + *pinst = inst;
>> + return 0;
>> +}
>> +
>> +void arch_optimize_kprobes(struct list_head *oplist)
>> +{
>> + struct optimized_kprobe *op, *tmp;
>> +
>> + list_for_each_entry_safe(op, tmp, oplist, list) {
>> + int err;
>> + unsigned int insn;
>> + WARN_ON(kprobe_disabled(&op->kp));
>> +
>> + /* Backup instructions which will be replaced by jump address */
>> + memcpy(op->optinsn.copied_insn, op->kp.addr, RELATIVEJUMP_SIZE);
>> +
>> + err = arm_branch_to_addr(&insn, op->kp.addr, op->optinsn.insn);
>> + BUG_ON(err);
>> +
>> + patch_text(op->kp.addr, insn);
>> +
>> + list_del_init(&op->list);
>> + }
>> + return;
>> +}
>> +
>> +void arch_unoptimize_kprobe(struct optimized_kprobe *op)
>> +{
>> + arch_arm_kprobe(&op->kp);
>> + return;
>> +}
>> +
>> +/*
>> + * Recover original instructions and breakpoints from relative jumps.
>> + * Caller must call with locking kprobe_mutex.
>> + */
>> +void arch_unoptimize_kprobes(struct list_head *oplist,
>> + struct list_head *done_list)
>> +{
>> + struct optimized_kprobe *op, *tmp;
>> +
>> + list_for_each_entry_safe(op, tmp, oplist, list) {
>> + arch_unoptimize_kprobe(op);
>> + list_move(&op->list, done_list);
>> + }
>> +}
>> +
>> +int arch_within_optimized_kprobe(struct optimized_kprobe *op,
>> + unsigned long addr)
>> +{
>> + return ((unsigned long)op->kp.addr <= addr &&
>> + (unsigned long)op->kp.addr + op->optinsn.size > addr);
>> +}
>> +
>> +void arch_remove_optimized_kprobe(struct optimized_kprobe *op)
>> +{
>> + __arch_remove_optimized_kprobe(op, 1);
>> +}
>>
>
>


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/