Message-Id: <20160909011138.b8e7add1364006fced47a877@kernel.org>
Date:   Fri, 9 Sep 2016 01:11:38 +0900
From:   Masami Hiramatsu <mhiramat@...nel.org>
To:     Anju T Sudhakar <anju@...ux.vnet.ibm.com>
Cc:     linux-kernel@...r.kernel.org, linuxppc-dev@...ts.ozlabs.org,
        ananth@...ibm.com, naveen.n.rao@...ux.vnet.ibm.com,
        paulus@...ba.org, srikar@...ux.vnet.ibm.com,
        benh@...nel.crashing.org, mpe@...erman.id.au,
        hemant@...ux.vnet.ibm.com, mahesh@...ux.vnet.ibm.com,
        mhiramat@...nel.org
Subject: Re: [PATCH 0/3] OPTPROBES for powerpc

Hi Anju,

On Wed,  7 Sep 2016 15:03:09 +0530
Anju T Sudhakar <anju@...ux.vnet.ibm.com> wrote:

> This is the patchset of the kprobes jump optimization
> (a.k.a OPTPROBES)for powerpc. Kprobe being an inevitable tool
> for kernel developers, enhancing the performance of kprobe has
> got much importance.
> 
> Currently kprobes inserts a trap instruction to probe a running kernel.
> Jump optimization allows kprobes to replace the trap with a branch,
> reducing the probe overhead drastically.

Thank you for updating the series :)
I'll check that.

> 
> In this series, conditional branch instructions are not considered for
> optimization as they have to be assessed carefully in SMP systems.

So, what kind of problem are there on PPC? (can conditional flag be
changed by other cpu?)

Thanks,

> 
> 
> Performance:
> =============
> An optimized kprobe in powerpc is 1.05 to 4.7 times faster than a kprobe.
> 
> Example:
> 
> Placed a probe at an offset 0x50 in _do_fork().
> *Time Diff here is, difference in time before hitting the probe and
> after the probed instruction. mftb() is employed in kernel/fork.c for
> this purpose.
> 
> # echo 0 > /proc/sys/debug/kprobes-optimization
> Kprobes globally unoptimized
> [  233.607120] Time Diff = 0x1f0
> [  233.608273] Time Diff = 0x1ee
> [  233.609228] Time Diff = 0x203
> [  233.610400] Time Diff = 0x1ec
> [  233.611335] Time Diff = 0x200
> [  233.612552] Time Diff = 0x1f0
> [  233.613386] Time Diff = 0x1ee
> [  233.614547] Time Diff = 0x212
> [  233.615570] Time Diff = 0x206
> [  233.616819] Time Diff = 0x1f3
> [  233.617773] Time Diff = 0x1ec
> [  233.618944] Time Diff = 0x1fb
> [  233.619879] Time Diff = 0x1f0
> [  233.621066] Time Diff = 0x1f9
> [  233.621999] Time Diff = 0x283
> [  233.623281] Time Diff = 0x24d
> [  233.624172] Time Diff = 0x1ea
> [  233.625381] Time Diff = 0x1f0
> [  233.626358] Time Diff = 0x200
> [  233.627572] Time Diff = 0x1ed
> 
> # echo 1 > /proc/sys/debug/kprobes-optimization
> Kprobes globally optimized
> [   70.797075] Time Diff = 0x103
> [   70.799102] Time Diff = 0x181
> [   70.801861] Time Diff = 0x15e
> [   70.803466] Time Diff = 0xf0
> [   70.804348] Time Diff = 0xd0
> [   70.805653] Time Diff = 0xad
> [   70.806477] Time Diff = 0xe0
> [   70.807725] Time Diff = 0xbe
> [   70.808541] Time Diff = 0xc3
> [   70.810191] Time Diff = 0xc7
> [   70.811007] Time Diff = 0xc0
> [   70.812629] Time Diff = 0xc0
> [   70.813640] Time Diff = 0xda
> [   70.814915] Time Diff = 0xbb
> [   70.815726] Time Diff = 0xc4
> [   70.816955] Time Diff = 0xc0
> [   70.817778] Time Diff = 0xcd
> [   70.818999] Time Diff = 0xcd
> [   70.820099] Time Diff = 0xcb
> [   70.821333] Time Diff = 0xf0
> 
> Implementation:
> ===================
> 
> The trap instruction is replaced by a branch to a detour buffer. To address
> the limitation of branch instruction in power architecture detour buffer
> slot is allocated from a reserved area . This will ensure that the branch
> is within ± 32 MB range. Patch 2/3 furnishes this. The current kprobes
> insn caches allocate memory area for insn slots with module_alloc(). This
> will always be beyond ± 32MB range.
> 
> The detour buffer contains a call to optimized_callback() which in turn
> call the pre_handler(). Once the pre-handler is run, the original
> instruction is emulated from the detour buffer itself. Also the detour
> buffer is equipped with a branch back to the normal work flow after the
> probed instruction is emulated. Before preparing optimization, Kprobes
> inserts original(breakpoint instruction)kprobe on the specified address.
> So, even if the kprobe is not possible to be optimized, it just uses a
> normal kprobe.
> 
> Limitations:
> ==============
> - Number of probes which can be optimized is limited by the size of the
>   area reserved.
> - Currently instructions which can be emulated are the only candidates for
>   optimization.
> - Conditional branch instructions are not optimized.
> - Probes on kernel module region are not considered for optimization now.
> 
> RFC patchset for optprobes: https://lkml.org/lkml/2016/5/31/375
> 			    https://lkml.org/lkml/2016/5/31/376
> 			    https://lkml.org/lkml/2016/5/31/377
> 			    https://lkml.org/lkml/2016/5/31/378 
> 
> Changes from RFC-v3 :
> 
> - Optimization for kporbe(in case of branch instructions) is limited to
>   unconditional branch instructions only, since the conditional
>   branches are to be assessed carefully in SMP systems.
> - create_return_branch() is omitted.
> - Comments by Masami are addressed.
>  
> 
> Anju T Sudhakar (3):
>   arch/powerpc : Add detour buffer support for optprobes
>   arch/powerpc : optprobes for powerpc core
>   arch/powerpc : Enable optprobes support in powerpc
> 
>  .../features/debug/optprobes/arch-support.txt      |   2 +-
>  arch/powerpc/Kconfig                               |   1 +
>  arch/powerpc/include/asm/kprobes.h                 |  24 ++
>  arch/powerpc/include/asm/sstep.h                   |   1 +
>  arch/powerpc/kernel/Makefile                       |   1 +
>  arch/powerpc/kernel/optprobes.c                    | 329 +++++++++++++++++++++
>  arch/powerpc/kernel/optprobes_head.S               | 119 ++++++++
>  arch/powerpc/lib/sstep.c                           |  21 ++
>  8 files changed, 497 insertions(+), 1 deletion(-)
>  create mode 100644 arch/powerpc/kernel/optprobes.c
>  create mode 100644 arch/powerpc/kernel/optprobes_head.S
> 
> -- 
> 2.7.4
> 


-- 
Masami Hiramatsu <mhiramat@...nel.org>

Powered by blists - more mailing lists