From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mga14.intel.com (mga14.intel.com [192.55.52.115]) by dpdk.org (Postfix) with ESMTP id 72C1E2904 for ; Mon, 1 Jan 2018 14:24:12 +0100 (CET) X-Amp-Result: SKIPPED(no attachment in message) X-Amp-File-Uploaded: False Received: from fmsmga006.fm.intel.com ([10.253.24.20]) by fmsmga103.fm.intel.com with ESMTP/TLS/DHE-RSA-AES256-GCM-SHA384; 01 Jan 2018 05:24:11 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.45,491,1508828400"; d="scan'208";a="191533615" Received: from dpdk-xiao-1.sh.intel.com ([10.67.110.153]) by fmsmga006.fm.intel.com with ESMTP; 01 Jan 2018 05:24:10 -0800 From: Xiao Wang To: ferruh.yigit@intel.com Cc: dev@dpdk.org, tiwei.bie@intel.com, stephen@networkplumber.org, Xiao Wang Date: Mon, 1 Jan 2018 14:00:10 -0800 Message-Id: <1514844010-39793-1-git-send-email-xiao.w.wang@intel.com> X-Mailer: git-send-email 1.8.3.1 In-Reply-To: <1513706293-133528-1-git-send-email-xiao.w.wang@intel.com> References: <1513706293-133528-1-git-send-email-xiao.w.wang@intel.com> Subject: [dpdk-dev] [PATCH v4] igb_uio: allow multi-process access 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: Mon, 01 Jan 2018 13:24:13 -0000 In some case, one device are accessed by different processes via different BARs, so one uio device may be opened by more than one process, for this case we just need to enable interrupt once, and pci_clear_master only when the last process closed. Fixes: 5f6ff30dc507 ("igb_uio: fix interrupt enablement after FLR in VM") Signed-off-by: Xiao Wang --- v4: - Make open/release function atomic, to ensure that other call won't return until the setup is really done. v3: - Use atomic_inc_return instead of atomic_inc and atomic_read. - Ditto for atomic_dec. v2: - Make uio reference counter operation atomic. --- lib/librte_eal/linuxapp/igb_uio/igb_uio.c | 18 ++++++++++++++++++ 1 file changed, 18 insertions(+) diff --git a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c index a3a98c1..1826adb 100644 --- a/lib/librte_eal/linuxapp/igb_uio/igb_uio.c +++ b/lib/librte_eal/linuxapp/igb_uio/igb_uio.c @@ -45,6 +45,8 @@ struct rte_uio_pci_dev { struct uio_info info; struct pci_dev *pdev; enum rte_intr_mode mode; + struct mutex lock; + int refcnt; }; static char *intr_mode; @@ -336,11 +338,18 @@ struct rte_uio_pci_dev { struct pci_dev *dev = udev->pdev; int err; + mutex_lock(&udev->lock); + if (++udev->refcnt > 1) { + mutex_unlock(&udev->lock); + return 0; + } + /* set bus master, which was cleared by the reset function */ pci_set_master(dev); /* enable interrupts */ err = igbuio_pci_enable_interrupts(udev); + mutex_unlock(&udev->lock); if (err) { dev_err(&dev->dev, "Enable interrupt fails\n"); return err; @@ -354,12 +363,19 @@ struct rte_uio_pci_dev { struct rte_uio_pci_dev *udev = info->priv; struct pci_dev *dev = udev->pdev; + mutex_lock(&udev->lock); + if (--udev->refcnt > 0) { + mutex_unlock(&udev->lock); + return 0; + } + /* disable interrupts */ igbuio_pci_disable_interrupts(udev); /* stop the device from further DMA */ pci_clear_master(dev); + mutex_unlock(&udev->lock); return 0; } @@ -480,6 +496,7 @@ struct rte_uio_pci_dev { if (!udev) return -ENOMEM; + mutex_init(&udev->lock); /* * enable device: ask low-level code to enable I/O and * memory @@ -570,6 +587,7 @@ struct rte_uio_pci_dev { { struct rte_uio_pci_dev *udev = pci_get_drvdata(dev); + mutex_destroy(&udev->lock); sysfs_remove_group(&dev->dev.kobj, &dev_attr_grp); uio_unregister_device(&udev->info); igbuio_pci_release_iomem(&udev->info); -- 1.8.3.1