From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga01.intel.com (mga01.intel.com [192.55.52.88]) by dpdk.org (Postfix) with ESMTP id 979A71BB7A; Tue, 3 Jul 2018 11:17:46 +0200 (CEST) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga001.fm.intel.com ([10.253.24.23]) by fmsmga101.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 03 Jul 2018 02:17:45 -0700 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.51,302,1526367600"; d="scan'208";a="69241403" Received: from aburakov-mobl.ger.corp.intel.com (HELO [10.237.220.109]) ([10.237.220.109]) by fmsmga001.fm.intel.com with ESMTP; 03 Jul 2018 02:17:44 -0700 To: Alejandro Lucero , dev@dpdk.org Cc: stable@dpdk.org, maxime.coquelin@redhat.com References: <1530552423-32301-1-git-send-email-alejandro.lucero@netronome.com> <1530552423-32301-5-git-send-email-alejandro.lucero@netronome.com> From: "Burakov, Anatoly" Message-ID: Date: Tue, 3 Jul 2018 10:17:43 +0100 User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:52.0) Gecko/20100101 Thunderbird/52.8.0 MIME-Version: 1.0 In-Reply-To: <1530552423-32301-5-git-send-email-alejandro.lucero@netronome.com> Content-Type: text/plain; charset=utf-8; format=flowed Content-Language: en-US Content-Transfer-Encoding: 7bit Subject: Re: [dpdk-stable] [PATCH 4/6] mem: use address hint for mapping hugepages X-BeenThere: stable@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: patches for DPDK stable branches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Tue, 03 Jul 2018 09:17:47 -0000 On 02-Jul-18 6:27 PM, Alejandro Lucero wrote: > Linux kernel uses a really high address as starting address for > serving mmaps calls. If there exists addressing limitations and > IOVA mode is VA, this starting address is likely too high for > those devices. However, it is possible to use a lower address in > the process virtual address space as with 64 bits there is a lot > of available space. > > This patch adds an address hint as starting address for 64 bits > systems. > > Signed-off-by: Alejandro Lucero > --- > long aligned_addr; > > - if (internal_config.base_virtaddr != 0) { > - addr = (void*) (uintptr_t) (internal_config.base_virtaddr + > - baseaddr_offset); > - } > - else addr = NULL; > - > RTE_LOG(DEBUG, EAL, "Ask a virtual area of 0x%zx bytes\n", *size); > > fd = open("/dev/zero", O_RDONLY); > @@ -278,7 +289,22 @@ > return NULL; > } > do { > - addr = mmap(addr, > + if (internal_config.base_virtaddr != 0) { > + addr_hint = (void *) (uintptr_t) > + (internal_config.base_virtaddr + > + baseaddr_offset); > + } > +#ifdef RTE_ARCH_64 > + else { > + addr_hint = (void *) (uintptr_t) (baseaddr + > + baseaddr_offset); > + } > +#else > + else { > + addr_hint = NULL; > + } > +#endif If my understanding is correct, calculations are all done on static variables, only the result is assigned to addr_hint which is local. Can we move this into a function and put these #ifdefs there, while keeping this code clean? > + addr = mmap(addr_hint, > (*size) + hugepage_sz, PROT_READ, > #ifdef RTE_ARCH_PPC_64 > MAP_PRIVATE | MAP_ANONYMOUS | MAP_HUGETLB, > @@ -286,8 +312,14 @@ > MAP_PRIVATE, > #endif > fd, 0); > - if (addr == MAP_FAILED) > + if (addr == MAP_FAILED) { > + /* map failed. Let's try with less memory */ > *size -= hugepage_sz; > + } else if (addr_hint && addr != addr_hint) { > + /* map not using hint. Let's try with another offset */ Comment is slightly misleading - "map not using hint" implies we are about to map something without using hint. Suggested rewording: suggested address hint was not used, try with another offset > + addr = MAP_FAILED; > + baseaddr_offset += 0x100000000; > + } > } while (addr == MAP_FAILED && *size > 0); > > if (addr == MAP_FAILED) { > -- Thanks, Anatoly