On Fri, 2019-02-01 at 19:28 +0000, Eads, Gage wrote: > > > > > -----Original Message----- > > From: Ola Liljedahl [mailto:ola.liljed...@arm.com] > > Sent: Friday, February 1, 2019 1:02 PM > > To: Eads, Gage <gage.e...@intel.com>; dev@dpdk.org > > Cc: jer...@marvell.com; chao...@linux.vnet.ibm.com; nd <n...@arm.com>; > > Richardson, Bruce <bruce.richard...@intel.com>; Ananyev, Konstantin > > <konstantin.anan...@intel.com>; hemant.agra...@nxp.com; > > olivier.m...@6wind.com; arybche...@solarflare.com; Gavin Hu (Arm > > Technology China) <gavin...@arm.com>; Honnappa Nagarahalli > > <honnappa.nagaraha...@arm.com> > > Subject: Re: [dpdk-dev] [PATCH 1/1] eal: add 128-bit cmpset (x86-64 only) > > > > On Fri, 2019-02-01 at 17:06 +0000, Eads, Gage wrote: > > > > > > > > > > > > > > > > > > -----Original Message----- > > > > From: Ola Liljedahl [mailto:ola.liljed...@arm.com] > > > > Sent: Monday, January 28, 2019 5:02 PM > > > > To: Eads, Gage <gage.e...@intel.com>; dev@dpdk.org > > > > Cc: arybche...@solarflare.com; jer...@marvell.com; > > > > chao...@linux.vnet.ibm.com; nd <n...@arm.com>; Richardson, Bruce > > > > <bruce.richard...@intel.com>; Ananyev, Konstantin > > > > <konstantin.anan...@intel.com>; hemant.agra...@nxp.com; > > > > olivier.m...@6wind.com; Honnappa Nagarahalli > > > > <honnappa.nagaraha...@arm.com>; Gavin Hu (Arm Technology China) > > > > <gavin...@arm.com> > > > > Subject: Re: [dpdk-dev] [PATCH 1/1] eal: add 128-bit cmpset (x86-64 > > > > only) > > > > > > > > On Mon, 2019-01-28 at 11:29 -0600, Gage Eads wrote: > > > > > > > > > > > > > > > This operation can be used for non-blocking algorithms, such as a > > > > > non-blocking stack or ring. > > > > > > > > > > Signed-off-by: Gage Eads <gage.e...@intel.com> > > > > > --- > > > > > .../common/include/arch/x86/rte_atomic_64.h | 31 > > > > > +++++++++++ > > > > > lib/librte_eal/common/include/generic/rte_atomic.h | 65 > > > > > ++++++++++++++++++++++ > > > > > 2 files changed, 96 insertions(+) > > > > > > > > > > diff --git > > > > > a/lib/librte_eal/common/include/arch/x86/rte_atomic_64.h > > > > > b/lib/librte_eal/common/include/arch/x86/rte_atomic_64.h > > > > > index fd2ec9c53..b7b90b83e 100644 > > > > > --- a/lib/librte_eal/common/include/arch/x86/rte_atomic_64.h > > > > > +++ b/lib/librte_eal/common/include/arch/x86/rte_atomic_64.h > > > > > @@ -34,6 +34,7 @@ > > > > > /* > > > > > * Inspired from FreeBSD src/sys/amd64/include/atomic.h > > > > > * Copyright (c) 1998 Doug Rabson > > > > > + * Copyright (c) 2019 Intel Corporation > > > > > * All rights reserved. > > > > > */ > > > > > > > > > > @@ -46,6 +47,7 @@ > > > > > > > > > > #include <stdint.h> > > > > > #include <rte_common.h> > > > > > +#include <rte_compat.h> > > > > > #include <rte_atomic.h> > > > > > > > > > > /*------------------------- 64 bit atomic operations > > > > > ------------------------ -*/ @@ -208,4 +210,33 @@ static inline > > > > > void rte_atomic64_clear(rte_atomic64_t *v) > > > > > } > > > > > #endif > > > > > > > > > > +static inline int __rte_experimental > > > > __rte_always_inline? > > > > > > > > > > > > > > > > > > > +rte_atomic128_cmpset(volatile rte_int128_t *dst, > > > > No need to declare the location volatile. Volatile doesn't do what > > > > you think it does. > > > > https://youtu.be/lkgszkPnV8g?t=1027 > > > > > > > I made this volatile to match the existing rte_atomicN_cmpset > > > definitions, which presumably have a good reason for using the > > > keyword. Maintainers, any input here? > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > + rte_int128_t *exp, > > > > I would declare 'exp' const as well and document that 'exp' is not > > > > updated (with the old value) for a failure. The reason being that > > > > ARMv8.0/AArch64 cannot atomically read the old value without also > > > > writing the location and that is bad for performance (unnecessary > > > > writes leads to unnecessary contention and worse scalability). And > > > > the user must anyway read the location (in the start of the critical > > > > section) using e.g. non-atomic 64-bit reads so there isn't actually > > > > any requirement for an atomic 128-bit read of the location. > > > > > > > Will change in v2. > > > > > > > > > > > > > > > > > > > > > > > > > > rte_int128_t *src, > > > > const rte_int128_t *src? > > > Sure, I don't see any harm in using const. > > > > > > > > > > > > > > > > > > > But why are we not passing 'exp' and 'src' by value? That works > > > > great, even with structs. Passing by value simplifies the compiler's > > > > life, especially if the call is inlined. Ask a compiler developer. > > > I ran objdump on the nb_stack code with both approaches, and > > > pass-by-reference resulted in fewer overall x86_64 assembly ops. > > > PBV: 100 ops for push, 97 ops for pop > > > PBR: 92 ops for push, 84 ops for pop > > OK I have never checked x86_64 code generation... I have good experiences > > with ARM/AArch64, everything seems to be done using registers. I am > > surprised > > there is a difference. > > > > Did a quick check with lfring, passing 'src' (third param) by reference and > > by > > value. No difference in code generation on x86_64. > > > > But if you insist let's go with PBR. > > > > > > > > > > > (Using the in-progress v5 nb_stack code) > > > > > > Another factor -- though much less compelling -- is that with pass-by- > > > reference, the user can create a 16B structure and cast it to > > > rte_int128_t when they call rte_atomic128_cmpset, whereas with > > > pass-by-value they need to put that struct in a union with rte_int128_t. > > Which is what I always do nowadays... Trying to use as few casts as possible > > and > > lie to the compiler as seldom as possible. But I can see the freedom > > provided by > > taking a pointer to something and cast it it rte_int128_t ptr in the call > > to rte_atomic128_cmpset(). > > > > Would prefer a name that is more similar to __atomic_compare_exchange(). > > E.g. > > rte_atomic128_compare_exchange() (or perhaps just rte_atomic128_cmpxchg)? > > All the rte_atomicXX_cmpset() functions do not take any memory order > > parameters. > > From an Arm perspective, we are not happy with that. > Since the function returns a boolean success value, isn't compare-and-set the > appropriate term? I was thinking of the memory ordering parameters that __atomic_xxx builtins have and we want (from an Arm perspective).
GCC __atomic_compare_exchange also returns a boolean. https://gcc.gnu.org/onlinedocs/gcc/_005f_005fatomic-Builtins.html bool __atomic_compare_exchange_n (type *ptr, type *expected, type desired, bool weak, int success_memorder, int failure_memorder) bool __atomic_compare_exchange (type *ptr, type *expected, type *desired, bool weak, int success_memorder, int failure_memorder) rte_atomic128_compare_exchange(rte_int128_t *dst, const rte_int128_t *exp, const rte_int182_t *src, bool weak, int mo_success, int mo_failure); > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > > + unsigned int weak, > > > > > + enum rte_atomic_memmodel_t success, > > > > > + enum rte_atomic_memmodel_t failure) { > > > > > + RTE_SET_USED(weak); > > > > > + RTE_SET_USED(success); > > > > > + RTE_SET_USED(failure); > > > > > + uint8_t res; > > > > > + > > > > > + asm volatile ( > > > > > + MPLOCKED > > > > > + "cmpxchg16b %[dst];" > > > > > + " sete %[res]" > > > > > + : [dst] "=m" (dst->val[0]), > > > > > + "=A" (exp->val[0]), > > > > > + [res] "=r" (res) > > > > > + : "c" (src->val[1]), > > > > > + "b" (src->val[0]), > > > > > + "m" (dst->val[0]), > > > > > + "d" (exp->val[1]), > > > > > + "a" (exp->val[0]) > > > > > + : "memory"); > > > > > + > > > > > + return res; > > > > > +} > > > > > + > > > > > #endif /* _RTE_ATOMIC_X86_64_H_ */ diff --git > > > > > a/lib/librte_eal/common/include/generic/rte_atomic.h > > > > > b/lib/librte_eal/common/include/generic/rte_atomic.h > > > > > index b99ba4688..8d612d566 100644 > > > > > --- a/lib/librte_eal/common/include/generic/rte_atomic.h > > > > > +++ b/lib/librte_eal/common/include/generic/rte_atomic.h > > > > > @@ -14,6 +14,7 @@ > > > > > > > > > > #include <stdint.h> > > > > > #include <rte_common.h> > > > > > +#include <rte_compat.h> > > > > > > > > > > #ifdef __DOXYGEN__ > > > > > > > > > > @@ -1082,4 +1083,68 @@ static inline void > > > > > rte_atomic64_clear(rte_atomic64_t > > > > > *v) > > > > > } > > > > > #endif > > > > > > > > > > +/*------------------------ 128 bit atomic operations > > > > > +------------------------ > > > > > -*/ > > > > > + > > > > > +/** > > > > > + * 128-bit integer structure. > > > > > + */ > > > > > +typedef struct { > > > > > + uint64_t val[2]; > > > > > +} __rte_aligned(16) rte_int128_t; > > > > So we can't use __int128? > > > > > > > I'll put it in a union with val[2], in case any implementations want > > > to use it. > > Thinking on this one more time, since the inline asm functions (e.g. for > > x86_64 > > cmpxchg16b and for AArch64 LDXP/STXP) anyway will use 64-bit registers, it > > makes most sense to make rte_int128_t a struct of 2x64b. The question is > > whether to use an array like above or a struct with two elements (which I > > normally do internally). Can you compare code generation with the following > > definition? > > typedef struct { > > uint64_t lo, hi; > > } __rte_aligned(16) rte_int128_t; > > > Interestingly, that made no difference in the PBV code but added more > instructions overall to PBR: > PBV: 100 insts for push, 97 insts for pop > PBR: 100 insts for push, 83 insts for pop I think we learned something here... Trying to understand exactly what. But I think this result settles it. I should test the different alternatives on Arm, does code generation behave the same as for x86_64. > > > > > > > > > > > > Thanks, > > > Gage > > > > > > [snip] > > -- > > Ola Liljedahl, Networking System Architect, Arm Phone +46706866373, Skype > > ola.liljedahl -- Ola Liljedahl, Networking System Architect, Arm Phone +46706866373, Skype ola.liljedahl