From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mail-lj1-f195.google.com (mail-lj1-f195.google.com [209.85.208.195]) by dpdk.org (Postfix) with ESMTP id 8201B1B3B8 for ; Fri, 29 Jun 2018 10:58:41 +0200 (CEST) Received: by mail-lj1-f195.google.com with SMTP id t22-v6so6693314ljc.11 for ; Fri, 29 Jun 2018 01:58:41 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=semihalf-com.20150623.gappssmtp.com; s=20150623; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc; bh=ojxf/S3GK47F6zERp5kJ8eoBGW87Hjvp1XPPKC4zdqE=; b=UvYZY74xPTYMYRZ+sRpFzDWWk9pbSx8lms6Jk6/nlqknj8FGfF0medoexXNVnFrtgB bCbpfBMJ87jJSqothYS+u8GVACtu8OqMngfqBMvpCwvbnQ80m53SNJKg8HhCsg10Nfwr cs8bpBIfMifY8YLkX8b2t4eA6X4rNrCdp2EjqNhCGbW427ZMXrZWpJeTBW87LJNz3Smv OUucllapg04YytKxdkJn1/eJRq9VjigbRJ44TNmQMQSE+BTYs67sl9WDA4tuQs+p7OjJ qDdfLcSdh7iZVw8Bj1YXLGYiXtLNmCUpvurmHNRTDx4efGv++8COd/e9h4joF/BJFaE5 GOiA== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:in-reply-to:references:from:date :message-id:subject:to:cc; bh=ojxf/S3GK47F6zERp5kJ8eoBGW87Hjvp1XPPKC4zdqE=; b=Kg8+yiskI12e3A69VsQKJKnQ4WF7YXX5VAHVi+Mr41ud8+5NELMxLgTgfUGNJFePv4 r8Sdd/I0onWa7ZvzUbSb7t/xBWZKHwLu5MKmaFWiqBCr9AbbzS/+tHIEsKVgdoJOZw6y o6ZDU5OfFa2/oNEGi7+R84DVzpCHJS3zuuXVkU5jE8UsZW7luIIYtPD9OdDkEFt1VNzp TtMvXmevZk78Lq+TGDAKfCz06Q/XqfZIpUT5YlRoXTYberYPvX/wAUasmWvIYsM0nUoR YbLxb3vLNs6vEB3yC+n26Nnpk7eMcHiYY/JSEIiYrunrkSYmHxsGpH7P+8cbbLlScXRI s17A== X-Gm-Message-State: APt69E3lOXiJeY+q6xCWVcJ4mx9vGZuFkru0Be3GdoUPNYQxfLpqE/DE i/8aq+V4VlvvpH1bXRXweFo1GuVueP83eb9lvpNucA== X-Google-Smtp-Source: ADUXVKKqZP53giY3fvfnFqv5sftt27YCo2OKua84Y04/Rax4UKkPlXP0FYQKrIAmP578x1dbaFJQsmJrDmnsi/uNWh0= X-Received: by 2002:a2e:750d:: with SMTP id q13-v6mr9284775ljc.56.1530262721091; Fri, 29 Jun 2018 01:58:41 -0700 (PDT) MIME-Version: 1.0 Received: by 2002:a19:4a95:0:0:0:0:0 with HTTP; Fri, 29 Jun 2018 01:58:40 -0700 (PDT) In-Reply-To: References: <1530191753-18689-1-git-send-email-rk@semihalf.com> <1530191753-18689-4-git-send-email-rk@semihalf.com> From: =?UTF-8?B?UmFmYcWCIEtvemlr?= Date: Fri, 29 Jun 2018 10:58:40 +0200 Message-ID: To: Ferruh Yigit Cc: dev@dpdk.org, Marcin Wojtas , =?UTF-8?Q?Micha=C5=82_Krawczyk?= , "Tzalik, Guy" , "Schmeilin, Evgeny" , "Matushevsky, Alexander" , "Chauskin, Igor" , Thomas Monjalon Content-Type: text/plain; charset="UTF-8" Subject: Re: [dpdk-dev] [PATCH v2 3/4] eal: enable WC during resources mapping X-BeenThere: dev@dpdk.org X-Mailman-Version: 2.1.15 Precedence: list List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 29 Jun 2018 08:58:41 -0000 2018-06-28 16:50 GMT+02:00 Ferruh Yigit : > On 6/28/2018 2:15 PM, Rafal Kozik wrote: >> Write combining (WC) increases NIC performance by making better >> utilization of PCI bus, but cannot be used by all PMDs. >> >> It will be enabled only if RTE_PCI_DRV_WC_ACTIVATE will be set in >> drivers flags. For proper work also igb_uio driver must be loaded with >> wc_activate set to 1. >> >> When mapping PCI resources, firstly try to us WC. >> If it is not supported it will fallback to normal mode. >> >> Signed-off-by: Rafal Kozik >> Acked-by: Bruce Richardson >> --- >> drivers/bus/pci/linux/pci_uio.c | 41 +++++++++++++++++++++++++++++------------ >> drivers/bus/pci/rte_bus_pci.h | 2 ++ >> 2 files changed, 31 insertions(+), 12 deletions(-) >> >> diff --git a/drivers/bus/pci/linux/pci_uio.c b/drivers/bus/pci/linux/pci_uio.c >> index d423e4b..e3947c2 100644 >> --- a/drivers/bus/pci/linux/pci_uio.c >> +++ b/drivers/bus/pci/linux/pci_uio.c >> @@ -282,22 +282,19 @@ int >> pci_uio_map_resource_by_index(struct rte_pci_device *dev, int res_idx, >> struct mapped_pci_resource *uio_res, int map_idx) >> { >> - int fd; >> + int fd = -1; >> char devname[PATH_MAX]; >> void *mapaddr; >> struct rte_pci_addr *loc; >> struct pci_map *maps; >> + int wc_activate = 0; >> + >> + if (dev->driver != NULL) >> + wc_activate = dev->driver->drv_flags & RTE_PCI_DRV_WC_ACTIVATE; >> >> loc = &dev->addr; >> maps = uio_res->maps; >> >> - /* update devname for mmap */ >> - snprintf(devname, sizeof(devname), >> - "%s/" PCI_PRI_FMT "/resource%d", >> - rte_pci_get_sysfs_path(), >> - loc->domain, loc->bus, loc->devid, >> - loc->function, res_idx); >> - >> /* allocate memory to keep path */ >> maps[map_idx].path = rte_malloc(NULL, strlen(devname) + 1, 0); >> if (maps[map_idx].path == NULL) { >> @@ -309,11 +306,31 @@ pci_uio_map_resource_by_index(struct rte_pci_device *dev, int res_idx, >> /* >> * open resource file, to mmap it >> */ >> - fd = open(devname, O_RDWR); >> - if (fd < 0) { >> - RTE_LOG(ERR, EAL, "Cannot open %s: %s\n", >> + if (wc_activate) { >> + /* update devname for mmap */ >> + snprintf(devname, sizeof(devname), >> + "%s/" PCI_PRI_FMT "/resource%d_wc", >> + rte_pci_get_sysfs_path(), >> + loc->domain, loc->bus, loc->devid, >> + loc->function, res_idx); >> + >> + fd = open(devname, O_RDWR); > > What do you think adding an error log here? If opening resource$_wc fails and > fallback to resource# file, there won't be any way for user to know about it. > This error will be misleading for two reasons: * even if open return success, it could silently fall-back to non prefetchable mode, * NICs could have multiple BARs, but not all support WC. I such case fails will be desirable. >> + } >> + >> + if (!wc_activate || fd < 0) { >> + snprintf(devname, sizeof(devname), >> + "%s/" PCI_PRI_FMT "/resource%d", >> + rte_pci_get_sysfs_path(), >> + loc->domain, loc->bus, loc->devid, >> + loc->function, res_idx); >> + >> + /* then try to map resource file */ >> + fd = open(devname, O_RDWR); >> + if (fd < 0) { >> + RTE_LOG(ERR, EAL, "Cannot open %s: %s\n", >> devname, strerror(errno)); >> - goto error; >> + goto error; >> + } >> } >> >> /* try mapping somewhere close to the end of hugepages */ >> diff --git a/drivers/bus/pci/rte_bus_pci.h b/drivers/bus/pci/rte_bus_pci.h >> index 458e6d0..828acc5 100644 >> --- a/drivers/bus/pci/rte_bus_pci.h >> +++ b/drivers/bus/pci/rte_bus_pci.h >> @@ -135,6 +135,8 @@ struct rte_pci_bus { >> >> /** Device needs PCI BAR mapping (done with either IGB_UIO or VFIO) */ >> #define RTE_PCI_DRV_NEED_MAPPING 0x0001 >> +/** Device needs PCI BAR mapping with enabled write combining (wc) */ >> +#define RTE_PCI_DRV_WC_ACTIVATE 0x0002 >> /** Device driver supports link state interrupt */ >> #define RTE_PCI_DRV_INTR_LSC 0x0008 >> /** Device driver supports device removal interrupt */ >> >