Re: [dpdk-dev] Fwd: [PATCH v3 2/2] eal/x86: Use lock-prefixed instructions to reduce cost of rte_smp_mb()

["Michael S. Tsirkin" <mst@redhat.com>]

On Mon, Jan 29, 2018 at 09:29:52AM +0000, Ananyev, Konstantin wrote:
> Hi Michael,
> 
> > 
> > On Mon, Jan 15, 2018 at 04:15:00PM +0100, Maxime Coquelin wrote:
> > > Hi Michael,
> > >
> > > FYI:
> > >
> > > -------- Forwarded Message --------
> > > Subject: [dpdk-dev] [PATCH v3 2/2] eal/x86: Use lock-prefixed instructions
> > > to reduce cost of rte_smp_mb()
> > > Date: Mon, 15 Jan 2018 15:09:31 +0000
> > > From: Konstantin Ananyev <konstantin.ananyev@intel.com>
> > > To: dev@dpdk.org
> > > CC: Konstantin Ananyev <konstantin.ananyev@intel.com>
> > >
> > > On x86 it  is possible to use lock-prefixed instructions to get
> > > the similar effect as mfence.
> > > As pointed by Java guys, on most modern HW that gives a better
> > > performance than using mfence:
> > > https://shipilev.net/blog/2014/on-the-fence-with-dependencies/
> > > That patch adopts that technique for rte_smp_mb() implementation.
> > > On BDW 2.2 mb_autotest on single lcore reports 2X cycle reduction,
> > > i.e. from ~110 to ~55 cycles per operation.
> > >
> > > Signed-off-by: Konstantin Ananyev <konstantin.ananyev@intel.com>
> > > Acked-by: Bruce Richardson <bruce.richardson@intel.com>
> > > ---
> > >  .../common/include/arch/x86/rte_atomic.h           | 44
> > > +++++++++++++++++++++-
> > >  1 file changed, 42 insertions(+), 2 deletions(-)
> > >
> > > diff --git a/lib/librte_eal/common/include/arch/x86/rte_atomic.h
> > > b/lib/librte_eal/common/include/arch/x86/rte_atomic.h
> > > index 8469f97e1..9d466d94a 100644
> > > --- a/lib/librte_eal/common/include/arch/x86/rte_atomic.h
> > > +++ b/lib/librte_eal/common/include/arch/x86/rte_atomic.h
> > > @@ -26,12 +26,52 @@ extern "C" {
> > >   #define	rte_rmb() _mm_lfence()
> > >  -#define rte_smp_mb() rte_mb()
> > > -
> > >  #define rte_smp_wmb() rte_compiler_barrier()
> > >   #define rte_smp_rmb() rte_compiler_barrier()
> > >  +/*
> > > + * From Intel Software Development Manual; Vol 3;
> > > + * 8.2.2 Memory Ordering in P6 and More Recent Processor Families:
> > > + * ...
> > > + * . Reads are not reordered with other reads.
> > > + * . Writes are not reordered with older reads.
> > > + * . Writes to memory are not reordered with other writes,
> > > + *   with the following exceptions:
> > > + *   . streaming stores (writes) executed with the non-temporal move
> > > + *     instructions (MOVNTI, MOVNTQ, MOVNTDQ, MOVNTPS, and MOVNTPD); and
> > > + *   . string operations (see Section 8.2.4.1).
> > > + *  ...
> > > + * . Reads may be reordered with older writes to different locations but
> > > not
> > > + * with older writes to the same location.
> > > + * . Reads or writes cannot be reordered with I/O instructions,
> > > + * locked instructions, or serializing instructions.
> > > + * . Reads cannot pass earlier LFENCE and MFENCE instructions.
> > > + * . Writes ... cannot pass earlier LFENCE, SFENCE, and MFENCE
> > > instructions.
> > > + * . LFENCE instructions cannot pass earlier reads.
> > > + * . SFENCE instructions cannot pass earlier writes ...
> > > + * . MFENCE instructions cannot pass earlier reads, writes ...
> > > + *
> > > + * As pointed by Java guys, that makes possible to use lock-prefixed
> > > + * instructions to get the same effect as mfence and on most modern HW
> > > + * that gives a better perfomance then using mfence:
> > > + * https://shipilev.net/blog/2014/on-the-fence-with-dependencies/
> > > + * Basic idea is to use lock prefixed add with some dummy memory location
> > > + * as the destination. From their experiments 128B(2 cache lines) below
> > > + * current stack pointer looks like a good candidate.
> > > + * So below we use that techinque for rte_smp_mb() implementation.
> > > + */
> > > +
> > > +static __rte_always_inline void
> > > +rte_smp_mb(void)
> > > +{
> > > +#ifdef RTE_ARCH_I686
> > > +	asm volatile("lock addl $0, -128(%%esp); " ::: "memory");
> > > +#else
> > > +	asm volatile("lock addl $0, -128(%%rsp); " ::: "memory");
> > > +#endif
> > > +}
> > > +
> > >  #define rte_io_mb() rte_mb()
> > >   #define rte_io_wmb() rte_compiler_barrier()
> > 
> > In my testing this appears to be suboptimal when the calling
> > function is large. The following seems to work better:
> > 
> > +static __rte_always_inline void
> > +rte_smp_mb(void)
> > +{
> > +#ifdef RTE_ARCH_I686
> > +	asm volatile("lock addl $0, -132(%%esp); " ::: "memory");
> > +#else
> > +	asm volatile("lock addl $0, -132(%%rsp); " ::: "memory");
> > +#endif
> > +}
> > +
> > 
> > The reason most likely is that 128 still overlaps the x86
> > red zone by 4 bytes.
> 
> I tried what you suggested but for my cases didn't see any improvement so far.
> Can you explain a bit more why do you expect it to be faster?
> Probably some particular scenario?
> Konstantin

It would depend on how much of a redzone is in use.
If the last 4 bytes of the red zone get used, you will
see a bit more stalls when trying to run the xor command.
You will need to call it from a function with lots of
scratch/temporary variables for that to be the case.

> > 
> > Feel free to reuse, and add
> > Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
> > 
> > 
> > > --
> > > 2.13.6