On Mon, Mar 25, 2019 at 02:26:27PM +0100, Gaëtan Rivet wrote: > On Mon, Mar 25, 2019 at 11:49:20AM +0000, Ferruh Yigit wrote: > > On 3/20/2019 11:46 AM, Thomas Monjalon wrote: > > > 20/03/2019 11:45, Ferruh Yigit: > > >> On 3/19/2019 8:30 PM, Thomas Monjalon wrote: > > >>> 19/03/2019 19:00, Ferruh Yigit: > > >>>> On 3/19/2019 5:36 PM, Thomas Monjalon wrote: > > >>>>> 19/03/2019 18:29, Ferruh Yigit: > > >>>>>> On 3/14/2019 10:04 PM, Thomas Monjalon wrote: > > >>>>>>> 14/03/2019 03:58, Hyong Youb Kim: > > >>>>>>>> On Wed, Mar 13, 2019 at 10:29:53PM +0100, Thomas Monjalon wrote: > > >>>>>>>>> 13/03/2019 22:11, John Daley (johndale): > > >>>>>>>>>> From: Thomas Monjalon > > >>>>>>>>>>> 13/03/2019 19:32, Ferruh Yigit: > > >>>>>>>>>>>> On 3/5/2019 7:11 AM, Hyong Youb Kim wrote: > > >>>>>>>>>>>>> The driver currently has a devarg to set the rewrite mode during > > >>>>>>>>>>>>> init. Some apps want to programatically set it after running > > >>>>>>>>>>>>> rte_eal_init() and finding that ports are VIC. Add a private > > >>>>>>>>>>>>> function to support such applications. > > >>>>>>>>>>>> > > >>>>>>>>>>>> It is not good idea to have PMD specific APIs (although we already have > > >>>>>>>>>>> some). > > >>>>>>>>>>>> > > >>>>>>>>>>>> Specific to this case, as far as I can see it is to pass a config > > >>>>>>>>>>>> value and do the action related to it, what would you think having a > > >>>>>>>>>>>> generic key/value set/get API in ethdev for this? Similar to rawdev > > >>>>>>>>>>> get_attr/set_attr [1]? > > >>>>>>>>>>>> > > >>>>>>>>>>>> My concern is it may turn into something like ioctl with many things > > >>>>>>>>>>>> pushed to it, and cause possible duplication ... > > >>>>>>>>>>> > > >>>>>>>>>>> Yes, it is clearly ioctl style. > > >>>>>>>>>>> > > >>>>>>>>>>> Please could you explain more what is the rewrite mode? > > >>>>>>>>>>> Does it apply to the port or the queue? > > >>>>>>>>>>> > > >>>>>>>>>> It applies to a port. By default the Cisco VIC VLAN tags every packet on ingress even if they were untagged coming in on the wire. They are tagged with VLAN 0 or a VLAN id programmed into the NIC depending on the configuration. Its part of the original design, to maintain priority bits, ancient history. > > >>>>>>>>>> > > >>>>>>>>>> Some apps don't like this (VPP) or take a slower path (OVS). Hyong added a ig-vlan-rewrite=untag devarg to disable this (leave untagged/default vlan packets untagged) during rte_eal_init and this is helpful for OVS, but VPP likes to set the rewrite mode after rte_eal_init() and finding the ports are VIC ports. So that is the reasoning behind the private API call. > > >>>>>>>>> > > >>>>>>>>> It looks like an application will always set this flag or never. > > >>>>>>>>> So I don't see the need for an API function. > > >>>>>>>>> Why VPP prefers set this flag later? > > >>>>>>>>> Would it be better to have some driver-specific flags, no matter the ports? > > >>>>>>>> > > >>>>>>>> As is, there seem to be two ways apps deal with NIC-specific tweaks/quirks. > > >>>>>>>> > > >>>>>>>> 1. Leave everything to the user. > > >>>>>>>> > > >>>>>>>> Let the human user specify NIC-specific settings (e.g. devarg, > > >>>>>>>> not-so-standard MTU/MRU, offloads with not-so-uniform behavior). The > > >>>>>>>> app simply passes these to DPDK and does nothing else. Devargs are > > >>>>>>>> passed to rte_eal_init. Other settings are applied during the > > >>>>>>>> configure phase after rte_eal_init. > > >>>>>>>> > > >>>>>>>> For example, OVS seems to go for a smallest common denominator that > > >>>>>>>> works with most NICs out of the box. Otherwiese, it kinda falls into > > >>>>>>>> this camp. > > >>>>>>>> > > >>>>>>>> For a problematic NIC that needs user intervention, troubleshooting > > >>>>>>>> goes like this :-) > > >>>>>>>> - Install app. > > >>>>>>>> - Run with settings that worked on a previous machine. > > >>>>>>>> - Some features suddenly do not work. > > >>>>>>>> - Google search this and that ("why this does not work on this server?"). > > >>>>>>>> - Contact vendors. > > >>>>>>>> - Find out this NIC has unexpected behavior. > > >>>>>>>> - Set devarg, tweak MTU/MRU, ... ("Oh need to set this for .."). > > >>>>>>>> - Now the features work. > > >>>>>>>> > > >>>>>>>> 2. Hide ugly tweaks from the user. > > >>>>>>>> > > >>>>>>>> VPP falls into this camp. The user specifies BDFs in the config (no > > >>>>>>>> devargs). The app calls rte_eal_init(BDFs), iterates through the > > >>>>>>>> discovered ports, applies whatever NIC-specific settings necessary > > >>>>>>>> during the configure phase (i.e. do this for vendor A NIC, do that for > > >>>>>>>> vendor B NIC), and then start the ports. > > >>>>>>>> > > >>>>>>>> The ingress vlan rewrite mode is devarg now, so is not usable in this > > >>>>>>>> model. One way around it is a private API. Driver specific flags > > >>>>>>>> during the configure phase would also work fine. Though, enic might be > > >>>>>>>> the only user of those flags. > > >>>>>>> > > >>>>>>> I think DPDK needs some driver configuration. > > >>>>>>> Currently the config is done per device with devargs. > > >>>>>>> The next devargs format allow this: > > >>>>>>> driver=enic,rewrite=on > > >>>>>>> and it can be passed to rte_eal_init(). > > >>>>>>> > > >>>>>>> We did not progress on the implementation of this format in recent months, > > >>>>>>> but you are welcome to help! > > >>>>>>> Instead of passing devargs in the whitelist/blacklist options, > > >>>>>>> we should introduce a new option, like --dev. > > >>>>>> > > >>>>>> But it will be still devarg in new implementation. > > >>>>> > > >>>>> With the new syntax, no need to specify a device. > > >>>>> We can match a driver or multiple drivers sharing the same property. > > >>>>> > > >>>>>> I guess for this use case, there is a need to pass this information from an API. > > >>>>>> Options can be: > > >>>>>> 1- PMD specific API > > >>>>>> 2- Extend ethdev dev_ops for each usecase > > >>>>>> 3- Have a generic API, as suggested above > > >>>>>> 4- Extend configure to accept flags > > >>>>>> > > >>>>>> I don't see a winner in above list, each has some problems. Any comment on how > > >>>>>> to proceed? > > >>>>> > > >>>>> I don't see a problem with the devargs approach. > > >>>> > > >>>> Devargs either passed via command line to DPDK application, or parameter to > > >>>> hotplug APIs. > > >>> > > >>> The application can pass whatever it wants to EAL. > > >> > > >> This means changing current device probe logic completely, right. > > >> Instead of probing everything on start, probe nothing and application add > > >> devices via eal (hotplug) API calls with providing devargs. > > >> I have no issue with this picture, only it doesn't look soon. > > > > > > No, I mean probe everything at startup automatically as usual. > > > Just need to pass an option to the driver > > > during its initialization. > > > > > >>> In the case described above, the application wants to enable > > >>> a mode of the driver for all its devices. > > >>> That's why I think the right solution is a driver option, > > >>> which can be achieved with the new devargs syntax. > > >>> > > >>>> If someone wants to use regular probe without any command line argument, and > > >>>> later configure the device via an API, can devargs be used? > > >>> > > >>> This is a scenario different of what is asked above. > > >>> In the case of a specific configuration of one device, > > >>> we have three choices. > > >>> These are your suggestions, with my comments: > > >>> 1- PMD specific API > > >>> 2- Extend ethdev dev_ops for each usecase > > >>> (3- Have a generic API) = choice 2 > > >>> (4- Extend configure to accept flags) = choice 1 > > >>> This is a choice 3: > > >>> - no support of exotic features > > >> > > >> Not sure if this is a real option for a vendor, HWs has exotic features and they > > >> want to enable it, I believe SW should not be the blocker for this. > > >> > > >> Also I definitely agree that exotic features should not break main/common usage, > > >> make it hard to use or make it confusing/complex etc. > > >> > > >> Until we have a better solution I guess we need to continue with private APIs. > > > > > > I think the driver option would work, > > > but it seems I fail to correctly explain it :) > > > > > > > I see it can work if an application always wants this config option to have > > *same* value. So it can set this in eal_init() always. > > > > This requires "driver=xxx,key=value" kind of support in devargs. > > > > > > John, Hyong, > > > > I guess some level of devargs support is already there, Thomas/Gaetan can > > provide more information on latest status of it, can it be possible to get some > > support from you to finalize this effort? > > > > And when it is ready enic can benefit from it for this usecase. > > > > Thanks, > > ferruh > > Hi Thomas, Ferruh, John, Hyong, > > driver=xxx,key=value could work, as it is not dependent on the > devargs framework, only on the driver implementation. Nothing specific > should be needed from EAL PoV (regarding this feature only). What will > be missing is the new devargs support in general. Sorry I spoke too quickly, specific development and some passing of arguments would be needed. > > Regarding the new devargs: this dev was stalled 2 versions ago at the > --dev inclusion step. This parameter was necessary for PMD maintainers > to be able to use the new init path with their drivers and transition to > rte_eth_devargs_parse() for devargs parsing. > > --dev was proposed, but its patch was not kept by Thomas during the > final crunch. I probably did not shout loud enough about it and let it > go, but I think this was a mistake: this feature was low-risk and central > in the transition process (it was in parallel to -w/b and --vdev). > > -- > Gaëtan Rivet > 6WIND -- Gaëtan Rivet 6WIND