From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id 1BDF643085; Wed, 16 Aug 2023 21:28:01 +0200 (CEST) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id 0D1EC411F3; Wed, 16 Aug 2023 21:28:01 +0200 (CEST) Received: from linux.microsoft.com (linux.microsoft.com [13.77.154.182]) by mails.dpdk.org (Postfix) with ESMTP id C4FEF40DDA for ; Wed, 16 Aug 2023 21:27:58 +0200 (CEST) Received: by linux.microsoft.com (Postfix, from userid 1086) id 28F25211F617; Wed, 16 Aug 2023 12:27:58 -0700 (PDT) DKIM-Filter: OpenDKIM Filter v2.11.0 linux.microsoft.com 28F25211F617 DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linux.microsoft.com; s=default; t=1692214078; bh=/5SFX6conqrXvRSfdBCP52Kp/n7qhT1jltSxXxsEW8Q=; h=Date:From:To:Cc:Subject:References:In-Reply-To:From; b=GwuHpWRpFosSq10oxCS7hbLRL3LnbtGvvwSJTII56v7/9Ho/JmRPpSxAavBidB2Vv WnMAWDLoUSjh27mZibGjI1IRxAo9sgCxN+gYsx4oq+jFOcqOiVS8vv/WiP9tKFAsK8 VHipsN5c0I4N7MoD4EeuKV0y2ckBIoJmd/9kLEDc= Date: Wed, 16 Aug 2023 12:27:58 -0700 From: Tyler Retzlaff To: Sivaprasad Tummala Cc: david.hunt@intel.com, anatoly.burakov@intel.com, ferruh.yigit@amd.com, david.marchand@redhat.com, thomas@monjalon.net, dev@dpdk.org Subject: Re: [PATCH v5 3/3] power: amd power monitor support Message-ID: <20230816192758.GA12453@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net> References: <20230418082529.544777-2-sivaprasad.tummala@amd.com> <20230816185959.1331336-1-sivaprasad.tummala@amd.com> <20230816185959.1331336-3-sivaprasad.tummala@amd.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20230816185959.1331336-3-sivaprasad.tummala@amd.com> User-Agent: Mutt/1.5.21 (2010-09-15) X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org On Wed, Aug 16, 2023 at 11:59:59AM -0700, Sivaprasad Tummala wrote: > mwaitx allows EPYC processors to enter a implementation dependent > power/performance optimized state (C1 state) for a specific period > or until a store to the monitored address range. > > Signed-off-by: Sivaprasad Tummala > Acked-by: Anatoly Burakov > --- > lib/eal/x86/rte_power_intrinsics.c | 77 +++++++++++++++++++++++++----- > 1 file changed, 66 insertions(+), 11 deletions(-) > > diff --git a/lib/eal/x86/rte_power_intrinsics.c b/lib/eal/x86/rte_power_intrinsics.c > index 6eb9e50807..b4754e17da 100644 > --- a/lib/eal/x86/rte_power_intrinsics.c > +++ b/lib/eal/x86/rte_power_intrinsics.c > @@ -17,6 +17,60 @@ static struct power_wait_status { > volatile void *monitor_addr; /**< NULL if not currently sleeping */ > } __rte_cache_aligned wait_status[RTE_MAX_LCORE]; > > +/** > + * These functions uses UMONITOR/UMWAIT instructions and will enter C0.2 state. > + * For more information about usage of these instructions, please refer to > + * Intel(R) 64 and IA-32 Architectures Software Developer's Manual. > + */ > +static void intel_umonitor(volatile void *addr) > +{ > + /* UMONITOR */ > + asm volatile(".byte 0xf3, 0x0f, 0xae, 0xf7;" > + : > + : "D"(addr)); > +} > + > +static void intel_umwait(const uint64_t timeout) > +{ > + const uint32_t tsc_l = (uint32_t)timeout; > + const uint32_t tsc_h = (uint32_t)(timeout >> 32); > + /* UMWAIT */ > + asm volatile(".byte 0xf2, 0x0f, 0xae, 0xf7;" > + : /* ignore rflags */ > + : "D"(0), /* enter C0.2 */ > + "a"(tsc_l), "d"(tsc_h)); > +} question and perhaps Anatoly Burakov can chime in with expertise. gcc/clang have built-in intrinsics for umonitor and umwait i believe as per our other thread of discussion is there a benefit to also providing inline assembly over just using the intrinsics? I understand that the intrinsics may not exist for the monitorx and mwaitx below so it is probably necessary for amd. so the suggestion here is when they are available just use the intrinsics. thanks > + > +/** > + * These functions uses MONITORX/MWAITX instructions and will enter C1 state. > + * For more information about usage of these instructions, please refer to > + * AMD64 Architecture Programmer's Manual. > + */ > +static void amd_monitorx(volatile void *addr) > +{ > + /* MONITORX */ > + asm volatile(".byte 0x0f, 0x01, 0xfa;" > + : > + : "a"(addr), > + "c"(0), /* no extensions */ > + "d"(0)); /* no hints */ > +} > + > +static void amd_mwaitx(const uint64_t timeout) > +{ > + /* MWAITX */ > + asm volatile(".byte 0x0f, 0x01, 0xfb;" > + : /* ignore rflags */ > + : "a"(0), /* enter C1 */ > + "c"(2), /* enable timer */ > + "b"(timeout)); > +} > + > +static struct { > + void (*mmonitor)(volatile void *addr); > + void (*mwait)(const uint64_t timeout); > +} __rte_cache_aligned power_monitor_ops; > + > static inline void > __umwait_wakeup(volatile void *addr) > { > @@ -75,8 +129,6 @@ int > rte_power_monitor(const struct rte_power_monitor_cond *pmc, > const uint64_t tsc_timestamp) > { > - const uint32_t tsc_l = (uint32_t)tsc_timestamp; > - const uint32_t tsc_h = (uint32_t)(tsc_timestamp >> 32); > const unsigned int lcore_id = rte_lcore_id(); > struct power_wait_status *s; > uint64_t cur_value; > @@ -109,10 +161,8 @@ rte_power_monitor(const struct rte_power_monitor_cond *pmc, > * versions support this instruction natively. > */ > > - /* set address for UMONITOR */ > - asm volatile(".byte 0xf3, 0x0f, 0xae, 0xf7;" > - : > - : "D"(pmc->addr)); > + /* set address for mmonitor */ > + power_monitor_ops.mmonitor(pmc->addr); > > /* now that we've put this address into monitor, we can unlock */ > rte_spinlock_unlock(&s->lock); > @@ -123,11 +173,8 @@ rte_power_monitor(const struct rte_power_monitor_cond *pmc, > if (pmc->fn(cur_value, pmc->opaque) != 0) > goto end; > > - /* execute UMWAIT */ > - asm volatile(".byte 0xf2, 0x0f, 0xae, 0xf7;" > - : /* ignore rflags */ > - : "D"(0), /* enter C0.2 */ > - "a"(tsc_l), "d"(tsc_h)); > + /* execute mwait */ > + power_monitor_ops.mwait(tsc_timestamp); > > end: > /* erase sleep address */ > @@ -173,6 +220,14 @@ RTE_INIT(rte_power_intrinsics_init) { > wait_multi_supported = 1; > if (i.power_monitor) > monitor_supported = 1; > + > + if (rte_cpu_get_flag_enabled(RTE_CPUFLAG_MONITORX)) { /* AMD */ > + power_monitor_ops.mmonitor = &amd_monitorx; > + power_monitor_ops.mwait = &amd_mwaitx; > + } else { /* Intel */ > + power_monitor_ops.mmonitor = &intel_umonitor; > + power_monitor_ops.mwait = &intel_umwait; > + } > } > > int > -- > 2.34.1