| Message ID | 54859A2C.8030009@huawei.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Mon, 2014-12-08 at 20:31 +0800, Wang Nan wrote: > On 2014/12/8 20:06, Wang Nan wrote: > > On 2014/12/8 19:50, Jon Medhurst (Tixy) wrote: > >> On Mon, 2014-12-08 at 19:15 +0800, Wang Nan wrote: > >>> On 2014/12/8 19:04, Jon Medhurst (Tixy) wrote: > >>>> On Mon, 2014-12-08 at 14:28 +0800, Wang Nan wrote: > [...] > >> > >> so another CPU could find and delete next before this one has finished > >> doing so. Would the list end up in a consistent state where no loops > >> develop and no probes are missed? I don't know the answer and a full > >> analysis would be complicated, but my gut feeling is that if a cpu can > >> observe the links in the list in an inconsistent state then only bad > >> things can result. > >> > > > > I see the problem. > > > > I'm thinking about making core.c and opt-arm.c to share stop_machine() code. > > stop_machine() is required when removing breakpoint, so I'd like to define > > a "remove_breakpoint" function in core.c and make opt-arm.c to call it. > > Do you think it is a good idea? > > > > > > What I mean is something like this: Yes, that should work, though as remove_breakpoint is a globally visible symbol, I suggest a less generic name for it, perhaps remove_kprobe_breakpoint ?
On 2014/12/8 21:22, Jon Medhurst (Tixy) wrote: > On Mon, 2014-12-08 at 20:31 +0800, Wang Nan wrote: >> On 2014/12/8 20:06, Wang Nan wrote: >>> On 2014/12/8 19:50, Jon Medhurst (Tixy) wrote: >>>> On Mon, 2014-12-08 at 19:15 +0800, Wang Nan wrote: >>>>> On 2014/12/8 19:04, Jon Medhurst (Tixy) wrote: >>>>>> On Mon, 2014-12-08 at 14:28 +0800, Wang Nan wrote: >> [...] >>>> >>>> so another CPU could find and delete next before this one has finished >>>> doing so. Would the list end up in a consistent state where no loops >>>> develop and no probes are missed? I don't know the answer and a full >>>> analysis would be complicated, but my gut feeling is that if a cpu can >>>> observe the links in the list in an inconsistent state then only bad >>>> things can result. >>>> >>> >>> I see the problem. >>> >>> I'm thinking about making core.c and opt-arm.c to share stop_machine() code. >>> stop_machine() is required when removing breakpoint, so I'd like to define >>> a "remove_breakpoint" function in core.c and make opt-arm.c to call it. >>> Do you think it is a good idea? >>> >>> >> >> What I mean is something like this: > > Yes, that should work, though as remove_breakpoint is a globally visible > symbol, I suggest a less generic name for it, perhaps > remove_kprobe_breakpoint ? > I don't think it is globally visible. Only files in arm/probes/kprobes can include "core.h". However, I do agree that remove_breakpoint() is not a good name. In my v15 patch, I'd like to rename it to kprobe_remove_breakpoint(), due to may of names defined in core.h are called kprobes_xxx. Thank you!
diff --git a/arch/arm/probes/kprobes/core.c b/arch/arm/probes/kprobes/core.c index 3a58db4..efd8ab1 100644 --- a/arch/arm/probes/kprobes/core.c +++ b/arch/arm/probes/kprobes/core.c @@ -163,19 +163,31 @@ void __kprobes arch_arm_kprobe(struct kprobe *p) * memory. It is also needed to atomically set the two half-words of a 32-bit * Thumb breakpoint. */ -int __kprobes __arch_disarm_kprobe(void *p) -{ - struct kprobe *kp = p; - void *addr = (void *)((uintptr_t)kp->addr & ~1); - - __patch_text(addr, kp->opcode); +struct patch { + void *addr; + unsigned int insn; +}; +static int __remove_breakpoint(void *data) +{ + struct patch *p = data; + __patch_text(p->addr, p->insn); return 0; } +void __kprobes remove_breakpoint(void *addr, unsigned int insn) +{ + struct patch p = { + .addr = addr, + .insn = insn, + }; + stop_machine(__remove_breakpoint, &p, cpu_online_mask); +} + void __kprobes arch_disarm_kprobe(struct kprobe *p) { - stop_machine(__arch_disarm_kprobe, p, cpu_online_mask); + remove_breakpoint((void *)((uintptr_t)p->addr & ~1), + p->opcode); } void __kprobes arch_remove_kprobe(struct kprobe *p) diff --git a/arch/arm/probes/kprobes/core.h b/arch/arm/probes/kprobes/core.h index f88c79f..7b7c334 100644 --- a/arch/arm/probes/kprobes/core.h +++ b/arch/arm/probes/kprobes/core.h @@ -30,6 +30,8 @@ #define KPROBE_THUMB16_BREAKPOINT_INSTRUCTION 0xde18 #define KPROBE_THUMB32_BREAKPOINT_INSTRUCTION 0xf7f0a018 +extern void remove_breakpoint(void *addr, unsigned int insn); + enum probes_insn __kprobes kprobe_decode_ldmstm(kprobe_opcode_t insn, struct arch_probes_insn *asi, const struct decode_header *h); diff --git a/arch/arm/probes/kprobes/opt-arm.c b/arch/arm/probes/kprobes/opt-arm.c index afbfeef..a1a1882 100644 --- a/arch/arm/probes/kprobes/opt-arm.c +++ b/arch/arm/probes/kprobes/opt-arm.c @@ -28,8 +28,9 @@ #include <asm/insn.h> /* for patch_text */ #include <asm/patch.h> -/* for stop_machine */ -#include <linux/stop_machine.h> + +#include "core.h" + /* * NOTE: the first sub and add instruction will be modified according * to the stack cost of the instruction. @@ -245,13 +246,8 @@ int arch_prepare_optimized_kprobe(struct optimized_kprobe *op, struct kprobe *or return 0; } -/* - * Similar to __arch_disarm_kprobe, operations which removing - * breakpoints must be wrapped by stop_machine to avoid racing. - */ -static __kprobes int __arch_optimize_kprobes(void *p) +void __kprobes arch_optimize_kprobes(struct list_head *oplist) { - struct list_head *oplist = p; struct optimized_kprobe *op, *tmp; list_for_each_entry_safe(op, tmp, oplist, list) { @@ -277,16 +273,15 @@ static __kprobes int __arch_optimize_kprobes(void *p) op->optinsn.copied_insn[0]) & 0xf0000000) | (insn & 0x0fffffff); - patch_text(op->kp.addr, insn); + /* + * Similar to __arch_disarm_kprobe, operations which + * removing breakpoints must be wrapped by stop_machine + * to avoid racing. + */ + remove_breakpoint(op->kp.addr, insn); list_del_init(&op->list); } - return 0; -} - -void arch_optimize_kprobes(struct list_head *oplist) -{ - stop_machine(__arch_optimize_kprobes, oplist, cpu_online_mask); } void arch_unoptimize_kprobe(struct optimized_kprobe *op)