DPDK patches and discussions
 help / color / mirror / Atom feed
From: Thomas Monjalon <thomas@monjalon.net>
To: "Xueming(Steven) Li" <xuemingl@nvidia.com>
Cc: dev@dpdk.org, Gaetan Rivet <gaetanr@nvidia.com>,
	"dev@dpdk.org" <dev@dpdk.org>, Asaf Penso <asafp@nvidia.com>,
	"david.marchand@redhat.com" <david.marchand@redhat.com>,
	"ferruh.yigit@intel.com" <ferruh.yigit@intel.com>,
	"andrew.rybchenko@oktetlabs.ru" <andrew.rybchenko@oktetlabs.ru>,
	"hemant.agrawal@nxp.com" <hemant.agrawal@nxp.com>,
	"stephen@networkplumber.org" <stephen@networkplumber.org>,
	"rosen.xu@intel.com" <rosen.xu@intel.com>,
	"ajit.khaparde@broadcom.com" <ajit.khaparde@broadcom.com>,
	"jerinj@marvell.com" <jerinj@marvell.com>
Subject: Re: [dpdk-dev] [PATCH v3 4/5] bus: add device arguments name parsing API
Date: Fri, 09 Apr 2021 01:49:43 +0200
Message-ID: <4292859.GXhFi8a5gR@thomas> (raw)
In-Reply-To: <BY5PR12MB432477520CD92E72BA4062DBA17B9@BY5PR12MB4324.namprd12.prod.outlook.com>

01/04/2021 17:13, Xueming(Steven) Li:
>From: Thomas Monjalon <thomas@monjalon.net>
> >30/03/2021 14:15, Xueming Li:
> >> To use Global Device Syntax as devargs, name is required for device
> >> management.
> >
> >Context is missing.
> >You mean the argument "name" for the vdev bus?
> 
> Devargs.name, it is used by probe and device iterator.

I think we could avoid having a name with the new syntax.
In my understanding, this is for compatibility with the legacy syntax.

> To locate a device from a devargs, devargs.name is compared
> agains name of probed devices.
> This not an issue for legacy syntax,
> since the string after "bus:" is saved as name.

It would be interesting to explain where the name is parsed
for the legacy syntax: rte_devargs_parse
and for the new syntax: rte_devargs_layers_parse called
in rte_devargs_parse in the next patch.

> >> In legacy parsing API, devargs name was extracted after bus name:
> >>   bus:name,kv_params,,,
> >>
> >> To parse new Global Device Syntax, this patch introduces new bus API
> >> to parse devargs and update name, different bus driver might choose
> >> different keys from parameters with unified formating, example:
> >>  -a bus=pci,addr=83:00.0/class=eth/driver=mlx5,...
> >>     name: 0000:03:00.0
> >>  -a bus=vdev,name=pcap0/class=eth/driver=pcap,...
> >>     name:pcap0
> >
> >Only PCI and vdev buses are implemented.
> >What can be the plan for others?
> >We should track the progress somewhere, maybe with TODO comments.
> 
> Like legacy parser, how about using "name" as default name key, the new syntax parser can resolve it by default.
> Then PCI bus overrides by using "addr" key in new bus API, 
> vdev and other bus drivers simply use default implementation, i.e. using "name" as key..

Yes, you mean if devargs_parse is not implemented by the bus driver,
the default is to parse the name property,
while the PCI implementation fills the devargs name with the addr property.

> >[...]
> >> +int
> >> +rte_pci_devargs_parse(struct rte_devargs *da) {
> >> +	struct rte_kvargs *kvargs;
> >> +	const char *addr_str;
> >> +	struct rte_pci_addr addr;
> >> +	int ret;
> >> +
> >> +	if (da == NULL)
> >> +		return 0;
> >> +	RTE_ASSERT(da->bus_str != NULL);
> >> +
> >> +	kvargs = rte_kvargs_parse(da->bus_str, NULL);
> >> +	if (kvargs == NULL) {
> >> +		RTE_LOG(ERR, EAL, "cannot parse argument list: %s\n",
> >> +			da->bus_str);
> >> +		ret = -ENODEV;
> >> +		goto out;
> >> +	}
> >> +
> >> +	addr_str = rte_kvargs_get(kvargs, pci_params_keys[RTE_PCI_PARAM_ADDR]);
> >> +	if (addr_str == NULL) {
> >> +		RTE_LOG(ERR, EAL, "No PCI address specified using '%s=<id>' in: %s\n",
> >> +			pci_params_keys[RTE_PCI_PARAM_ADDR], da->bus_str);
> >> +		ret = -ENODEV;
> >> +		goto out;
> >> +	}
> >> +
> >> +	ret = rte_pci_addr_parse(addr_str, &addr);
> >> +	if (ret != 0) {
> >> +		RTE_LOG(ERR, EAL, "PCI address invalid: %s\n", da->bus_str);
> >> +		ret = -EINVAL;
> >> +		goto out;
> >> +	}
> >> +
> >> +	rte_pci_device_name(&addr, da->name, sizeof(da->name));
> >> +
> >> +	/* TODO: class parse -> driver parse */
> >
> >Please could you give a longer explanation of what is missing?
> 
> Just an option to give class driver and device driver to override name parsing.
> But this should happen in new devargs parser, not here if needed.
> 
> I'll remove this line, and mention it in commit log.

What would be the benefit of overriding devargs name parsing?
I feel it would be better to not rely on the devargs name
with the new syntax if we have such need.



  reply	other threads:[~2021-04-08 23:49 UTC|newest]

Thread overview: 119+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-12-18 15:16 [dpdk-dev] [RFC 0/9] support global syntax Xueming Li
2020-12-18 15:16 ` [dpdk-dev] [RFC 1/9] devargs: fix data buffer storage type Xueming Li
2021-01-08 14:54   ` [dpdk-dev] [PATCH v1 0/7] eal: support global syntax Xueming Li
2021-01-08 14:54   ` [dpdk-dev] [PATCH v1 1/7] devargs: fix data buffer storage type Xueming Li
2021-01-08 14:54   ` [dpdk-dev] [PATCH v1 2/7] devargs: fix memory leak on parsing error Xueming Li
2021-01-08 14:54   ` [dpdk-dev] [PATCH v1 3/7] devargs: fix memory leak in legacy parser Xueming Li
2021-01-08 14:54   ` [dpdk-dev] [PATCH v1 4/7] devargs: fix buffer data memory leak Xueming Li
2021-01-08 14:54   ` [dpdk-dev] [PATCH v1 5/7] kvargs: add get by key function Xueming Li
2021-01-08 14:54   ` [dpdk-dev] [PATCH v1 6/7] devargs: support new global device syntax Xueming Li
2021-01-08 14:54   ` [dpdk-dev] [PATCH v1 7/7] bus/pci: add new global device syntax support Xueming Li
2021-01-08 15:14   ` [dpdk-dev] [PATCH v1 0/2] mlx5: support global syntax Xueming Li
2021-01-08 15:14   ` [dpdk-dev] [PATCH v1 1/2] common/mlx5: support device " Xueming Li
2021-01-08 15:15   ` [dpdk-dev] [PATCH v1 2/2] net/mlx5: support new " Xueming Li
2021-01-18 11:29   ` [dpdk-dev] [PATCH v3 0/9] net/mlx5: support SubFunction representor Xueming Li
2021-01-18 11:29   ` [dpdk-dev] [PATCH v3 1/9] common/mlx5: update representor name parsing Xueming Li
2021-01-18 11:29   ` [dpdk-dev] [PATCH v3 2/9] net/mlx5: support representor of sub function Xueming Li
2021-01-18 11:29   ` [dpdk-dev] [PATCH v3 3/9] net/mlx5: revert setting bonding representor to first PF Xueming Li
2021-01-18 11:29   ` [dpdk-dev] [PATCH v3 4/9] net/mlx5: refactor bonding representor probe Xueming Li
2021-01-18 11:29   ` [dpdk-dev] [PATCH v3 5/9] net/mlx5: support representor from multiple PFs Xueming Li
2021-01-18 11:29   ` [dpdk-dev] [PATCH v3 6/9] net/mlx5: save bonding member ports information Xueming Li
2021-01-18 11:29   ` [dpdk-dev] [PATCH v3 7/9] " Xueming Li
2021-01-18 16:17     ` Slava Ovsiienko
2021-01-18 23:05       ` Xueming(Steven) Li
2021-01-18 11:29   ` [dpdk-dev] [PATCH v3 8/9] net/mlx5: fix setting VF default MAC through representor Xueming Li
2021-01-18 11:29   ` [dpdk-dev] [PATCH v3 9/9] net/mlx5: improve bonding xstats Xueming Li
2021-01-18 15:16   ` [dpdk-dev] [PATCH v2 0/5] eal: enable global device syntax Xueming Li
2021-01-18 15:16   ` [dpdk-dev] [PATCH v2 1/5] devargs: fix memory leak on parsing error Xueming Li
2021-03-18  9:12     ` Thomas Monjalon
2021-01-18 15:16   ` [dpdk-dev] [PATCH v2 2/5] devargs: refactor scratch buffer storage Xueming Li
2021-03-18  9:14     ` Thomas Monjalon
2021-01-18 15:16   ` [dpdk-dev] [PATCH v2 3/5] kvargs: add get by key function Xueming Li
2021-01-18 15:16   ` [dpdk-dev] [PATCH v2 4/5] devargs: parse name from global device syntax Xueming Li
2021-01-18 15:16   ` [dpdk-dev] [PATCH v2 5/5] devargs: enable global device syntax devargs Xueming Li
2021-01-18 15:26   ` [dpdk-dev] [PATCH v2 0/2] mlx5: support global device syntax Xueming Li
2021-01-18 15:26   ` [dpdk-dev] [PATCH v2 1/2] common/mlx5: support device global syntax Xueming Li
2021-04-05 10:54     ` Slava Ovsiienko
2021-04-08 12:24       ` Raslan Darawsheh
2021-04-08 14:04         ` Raslan Darawsheh
2021-04-08 14:08           ` Xueming(Steven) Li
2021-04-08 14:13             ` Raslan Darawsheh
2021-04-19  9:29               ` Raslan Darawsheh
2021-04-19 10:36                 ` Raslan Darawsheh
2021-01-18 15:26   ` [dpdk-dev] [PATCH v2 2/2] net/mlx5: support new global device syntax Xueming Li
2021-04-05 10:56     ` Slava Ovsiienko
2021-01-19  7:28   ` [dpdk-dev] [PATCH v4 0/8] net/mlx5: support SubFunction representor Xueming Li
2021-01-19  7:28   ` [dpdk-dev] [PATCH v4 1/8] common/mlx5: update representor name parsing Xueming Li
2021-01-19  7:28   ` [dpdk-dev] [PATCH v4 2/8] net/mlx5: support representor of sub function Xueming Li
2021-01-19  7:28   ` [dpdk-dev] [PATCH v4 3/8] net/mlx5: revert setting bonding representor to first PF Xueming Li
2021-01-19  7:28   ` [dpdk-dev] [PATCH v4 4/8] net/mlx5: refactor bonding representor probe Xueming Li
2021-01-19  7:28   ` [dpdk-dev] [PATCH v4 5/8] net/mlx5: support representor from multiple PFs Xueming Li
2021-01-19  7:28   ` [dpdk-dev] [PATCH v4 6/8] net/mlx5: save bonding member ports information Xueming Li
2021-01-19  7:28   ` [dpdk-dev] [PATCH v4 7/8] net/mlx5: fix setting VF default MAC through representor Xueming Li
2021-01-19  7:28   ` [dpdk-dev] [PATCH v4 8/8] net/mlx5: improve bonding xstats Xueming Li
2021-03-28 13:48   ` [dpdk-dev] [PATCH v5 0/9] net/mlx5: support SubFunction representor Xueming Li
2021-03-31  7:20     ` Raslan Darawsheh
2021-03-31  7:27       ` Xueming(Steven) Li
2021-03-28 13:48   ` [dpdk-dev] [PATCH v5 1/9] common/mlx5: sub-function representor port name parsing Xueming Li
2021-03-28 13:48   ` [dpdk-dev] [PATCH v5 2/9] net/mlx5: support representor of sub function Xueming Li
2021-03-28 13:48   ` [dpdk-dev] [PATCH v5 3/9] net/mlx5: revert setting bonding representor to first PF Xueming Li
2021-03-28 13:48   ` [dpdk-dev] [PATCH v5 4/9] net/mlx5: refactor bonding representor probe Xueming Li
2021-03-28 13:48   ` [dpdk-dev] [PATCH v5 5/9] net/mlx5: support list value of representor PF Xueming Li
2021-03-28 13:48   ` [dpdk-dev] [PATCH v5 6/9] net/mlx5: save bonding member ports information Xueming Li
2021-03-28 13:48   ` [dpdk-dev] [PATCH v5 7/9] net/mlx5: fix setting VF default MAC through representor Xueming Li
2021-03-31  7:46     ` Raslan Darawsheh
2021-03-28 13:48   ` [dpdk-dev] [PATCH v5 8/9] net/mlx5: improve xstats of bonding port Xueming Li
2021-03-28 13:48   ` [dpdk-dev] [PATCH v5 9/9] net/mlx5: probe host PF representor with SubFunction Xueming Li
2021-03-30  7:37     ` Slava Ovsiienko
2021-03-30 12:15   ` [dpdk-dev] [PATCH v3 0/5] eal: enable global device syntax by default Xueming Li
2021-03-31  8:23     ` Gaëtan Rivet
2021-03-30 12:15   ` [dpdk-dev] [PATCH v3 1/5] devargs: unify scratch buffer storage Xueming Li
2021-04-01  9:04     ` Kinsella, Ray
2021-03-30 12:15   ` [dpdk-dev] [PATCH v3 2/5] devargs: fix memory leak on parsing error Xueming Li
2021-03-30 12:15   ` [dpdk-dev] [PATCH v3 3/5] kvargs: add get by key function Xueming Li
2021-04-01  9:06     ` Kinsella, Ray
2021-04-01  9:10       ` Xueming(Steven) Li
2021-03-30 12:15   ` [dpdk-dev] [PATCH v3 4/5] bus: add device arguments name parsing API Xueming Li
2021-03-31 10:19     ` Thomas Monjalon
2021-04-01 15:13       ` Xueming(Steven) Li
2021-04-08 23:49         ` Thomas Monjalon [this message]
2021-03-30 12:15   ` [dpdk-dev] [PATCH v3 5/5] devargs: parse global device syntax Xueming Li
2021-04-10 14:23   ` [dpdk-dev] [PATCH v4 0/5] eal: enable global device syntax by default Xueming Li
2021-04-10 14:23   ` [dpdk-dev] [PATCH v4 1/5] devargs: unify scratch buffer storage Xueming Li
2021-04-10 19:59     ` Tal Shnaiderman
2021-04-12 12:07       ` Xueming(Steven) Li
2021-04-10 14:23   ` [dpdk-dev] [PATCH v4 2/5] devargs: fix memory leak on parsing error Xueming Li
2021-04-10 14:23   ` [dpdk-dev] [PATCH v4 3/5] kvargs: add get by key function Xueming Li
2021-04-12  6:52     ` Olivier Matz
2021-04-12 12:07       ` Xueming(Steven) Li
2021-04-12 21:18         ` Thomas Monjalon
2021-04-10 14:23   ` [dpdk-dev] [PATCH v4 4/5] bus: add device arguments name parsing API Xueming Li
2021-04-12 21:16     ` Thomas Monjalon
2021-04-12 23:37       ` Xueming(Steven) Li
2021-04-10 14:23   ` [dpdk-dev] [PATCH v4 5/5] devargs: parse global device syntax Xueming Li
2021-04-12 21:24     ` Thomas Monjalon
2021-04-12 23:47       ` Xueming(Steven) Li
2021-04-13  3:14   ` [dpdk-dev] [PATCH v5 0/5] eal: enable global device syntax by default Xueming Li
2021-04-14 19:49     ` Thomas Monjalon
2021-04-23 11:06       ` Kinsella, Ray
2021-04-23 11:39         ` Gaëtan Rivet
2021-04-23 12:35           ` Kinsella, Ray
2021-09-02 14:46       ` Thomas Monjalon
2021-04-13  3:14   ` [dpdk-dev] [PATCH v5 1/5] devargs: unify scratch buffer storage Xueming Li
2021-04-16  7:00     ` David Marchand
2021-04-16 12:32       ` Aaron Conole
2021-04-16 12:43         ` [dpdk-dev] [dpdklab] " Lincoln Lavoie
2021-04-16 12:58           ` Thomas Monjalon
2021-04-16 13:14             ` Lincoln Lavoie
2021-04-13  3:14   ` [dpdk-dev] [PATCH v5 2/5] devargs: fix memory leak on parsing error Xueming Li
2021-04-13  3:14   ` [dpdk-dev] [PATCH v5 3/5] kvargs: add get by key function Xueming Li
2021-04-13  3:14   ` [dpdk-dev] [PATCH v5 4/5] bus: add device arguments name parsing API Xueming Li
2021-04-13  3:14   ` [dpdk-dev] [PATCH v5 5/5] devargs: parse global device syntax Xueming Li
2020-12-18 15:16 ` [dpdk-dev] [RFC 2/9] devargs: fix memory leak on parsing error Xueming Li
2020-12-18 15:16 ` [dpdk-dev] [RFC 3/9] devargs: fix memory leak in legacy parser Xueming Li
2020-12-18 15:16 ` [dpdk-dev] [RFC 4/9] devargs: fix buffer data memory leak Xueming Li
2020-12-18 15:16 ` [dpdk-dev] [RFC 5/9] kvargs: add get by key function Xueming Li
2020-12-18 15:16 ` [dpdk-dev] [RFC 6/9] devargs: support new global device syntax Xueming Li
2020-12-18 15:16 ` [dpdk-dev] [RFC 7/9] bus/pci: add new global device syntax support Xueming Li
2020-12-18 15:16 ` [dpdk-dev] [RFC 8/9] common/mlx5: support device global syntax Xueming Li
2020-12-18 15:16 ` [dpdk-dev] [RFC 9/9] net/mlx5: support new " Xueming Li

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=4292859.GXhFi8a5gR@thomas \
    --to=thomas@monjalon.net \
    --cc=ajit.khaparde@broadcom.com \
    --cc=andrew.rybchenko@oktetlabs.ru \
    --cc=asafp@nvidia.com \
    --cc=david.marchand@redhat.com \
    --cc=dev@dpdk.org \
    --cc=ferruh.yigit@intel.com \
    --cc=gaetanr@nvidia.com \
    --cc=hemant.agrawal@nxp.com \
    --cc=jerinj@marvell.com \
    --cc=rosen.xu@intel.com \
    --cc=stephen@networkplumber.org \
    --cc=xuemingl@nvidia.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

DPDK patches and discussions

This inbox may be cloned and mirrored by anyone:

	git clone --mirror https://inbox.dpdk.org/dev/0 dev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 dev dev/ https://inbox.dpdk.org/dev \
		dev@dpdk.org
	public-inbox-index dev

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.dev


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git