From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Marchand Subject: [PATCH v4 00/10] kill global pci device id list (almost) Date: Mon, 11 Jul 2016 16:40:35 +0200 Message-ID: <1468248045-3495-1-git-send-email-david.marchand@6wind.com> References: <1461156236-25349-1-git-send-email-david.marchand@6wind.com> Cc: bruce.richardson@intel.com, thomas.monjalon@6wind.com, nhorman@tuxdriver.com, pmatilai@redhat.com, christian.ehrhardt@canonical.com, stephen@networkplumber.com, Wenzhuo Lu , Helin Zhang , Jingjing Wu , Jing Chen , Huawei Xie , Yuanhan Liu , Yong Wang , John Daley , Nelson Escobar , Sony Chacko , Harish Patil , Rasesh Mody , Stephen Hurd , Jan Medala , Jakub Palider , Netanel Belgazal , Evgeny Schemeilin To: dev@dpdk.org Return-path: Received: from mail-wm0-f54.google.com (mail-wm0-f54.google.com [74.125.82.54]) by dpdk.org (Postfix) with ESMTP id 046F32B98 for ; Mon, 11 Jul 2016 16:40:58 +0200 (CEST) Received: by mail-wm0-f54.google.com with SMTP id f126so93303156wma.1 for ; Mon, 11 Jul 2016 07:40:58 -0700 (PDT) In-Reply-To: <1461156236-25349-1-git-send-email-david.marchand@6wind.com> List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" With the introduction of pmdinfo by Neil, we have almost everything in place to get rid of the pci devices in eal. We still have some ties with some pmds for functionalities like kni/ethtool or ixgbe bypass api, so the plan has switched to touch all pmds but those igb and ixgbe drivers. Since we still need rte_pci_dev_ids.h for those drivers, I just stripped the doxygen parts to stop referencing it in the documentation. I have validated this patchset by comparing the pmdinfo outputs and just noticed a difference for enic (where the pci ids were registered twice before). Yet, please maintainers review carefully. Thanks. Changes since v3: - dropped my approach at extracting informations from binaries - let igb{,vf} and ixgbe{,vf} untouched - rebased on HEAD - added bnxt Changes since v2: - rebased on HEAD - ena driver has been aligned - this patchset now depends on [1] as it avoids touching all drivers this way - not storing the pci ids in a dedicated section anymore, pci drivers are exported and parsed by a quickly written (and naive) tool Changes since v1: - indent fixes in i40e, fm10k, virtio, vmxnet3, enic, bnx2x. - rebased on head (ixgbe update) - removed doc update (will be sent separately) [1]: http://dpdk.org/ml/archives/dev/2016-April/037686.html -- David Marchand David Marchand (10): eal: remove PCI device ids header from doxygen net/e1000: move em PCI device ids to the driver net/i40e: move PCI device ids to the driver net/fm10k: move PCI device ids to the driver net/virtio: move PCI device ids to the driver net/vmxnet3: move PCI device ids to the driver net/enic: move PCI device ids to the driver net/bnx2x: move PCI device ids to the driver net/bnxt: move PCI device ids to the driver net/ena: remove unneeded pci macro doc/api/doxy-api-index.md | 1 - drivers/net/bnx2x/bnx2x.c | 3 +- drivers/net/bnx2x/bnx2x_ethdev.c | 21 +- drivers/net/bnxt/bnxt_ethdev.c | 27 +- drivers/net/e1000/e1000_ethdev.h | 2 + drivers/net/e1000/em_ethdev.c | 37 ++- drivers/net/ena/ena_ethdev.c | 7 +- drivers/net/enic/enic_ethdev.c | 12 +- drivers/net/fm10k/fm10k_ethdev.c | 6 +- drivers/net/i40e/i40e_ethdev.c | 25 +- drivers/net/i40e/i40e_ethdev_vf.c | 9 +- drivers/net/virtio/virtio_ethdev.c | 7 +- drivers/net/vmxnet3/vmxnet3_ethdev.c | 9 +- lib/librte_eal/common/include/rte_pci_dev_ids.h | 420 ------------------------ 14 files changed, 120 insertions(+), 466 deletions(-) -- 1.9.1