From mboxrd@z Thu Jan 1 00:00:00 1970 From: Aaron Conole Subject: Re: [dpdk-stable] [RFC 2/2] nfp: allow for non-root user Date: Fri, 20 Apr 2018 10:56:06 -0400 Message-ID: References: <20180412222208.11770-1-aconole@redhat.com> <20180412222208.11770-3-aconole@redhat.com> <5d0e69b9-3655-f3f6-9cbb-1090b9471813@intel.com> Mime-Version: 1.0 Content-Type: text/plain Cc: Alejandro Lucero , dev , Adrien Mazarguil , stable@dpdk.org, Thomas Monjalon To: Ferruh Yigit Return-path: In-Reply-To: <5d0e69b9-3655-f3f6-9cbb-1090b9471813@intel.com> (Ferruh Yigit's message of "Fri, 20 Apr 2018 15:12:28 +0100") List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Ferruh Yigit writes: > On 4/19/2018 7:05 AM, Alejandro Lucero wrote: >> On Wed, Apr 18, 2018 at 1:32 PM, Aaron Conole wrote: >> >>> Alejandro Lucero writes: >>> >>>> On Tue, Apr 17, 2018 at 8:19 PM, Aaron Conole >>> wrote: >>>> >>>> Alejandro Lucero writes: >>>> >>>> > I was just wondering, if device device PCI sysfs resource files or >>> VFIO group /dev files >>>> require to change >>>> > permissions for non-root users, does it not make sense to adjust also >>> /var/lock in the >>>> system? >>>> >>>> For the /dev, we use udev rules - so the correct individual vfio device >>>> files get assigned the correct permissions. No such mechanism exists >>>> for /var/lock as far as I can tell. >>>> >>>> Ex. see: >>>> >>>> https://github.com/openvswitch/ovs/blob/master/ >>> rhel/usr_lib_udev_rules.d_91-vfio.rules >>>> >>>> >>>> Maybe something similar exists that we could use to generate the lock >>>> file automatically? >>>> >>>> What about /sysfs/bus/pci/device/$PCI_DEV/resource file? >>>> >>>> Is RH forcing OVS DPDK to only work if the host has IOMMU support? >>> >>> Yes. >>> >> >> Ok then. It makes sense now to apply this patch to stable versions. >> >> Acked-by: Alejandro Lucero > > Since the target is the stable tree, I will drop them from patchwork as not > applicable. > > Can you please send v1 of the patch to the stable mail list, it can be good idea > to cc stable maintainers as well. Will do. I'll spin into a proper series and submit next week. Sorry for the extra noise, Ferruh.