From: Ed Czeck <ed.czeck@atomicrules.com>
To: Ferruh Yigit <ferruh.yigit@intel.com>
Cc: dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH v4 1/7] net/ark: PMD for Atomic Rules Arkville driver stub
Date: Thu, 23 Mar 2017 15:46:13 -0400 [thread overview]
Message-ID: <CALZ3GuhCCVPFJ9kqxgnecQTNcf9M9uqVxMN+mOJR5VKckScZwA@mail.gmail.com> (raw)
In-Reply-To: <0d1c21b6-6163-12f0-cbe4-d23c06a809e9@intel.com>
In the next patch set, v5 the following are addressed.
>
>
> > +CONFIG_RTE_LIBRTE_ARK_PMD=n
> Is it not tested or known that it is not supported?
> > CONFIG_RTE_LIBRTE_IXGBE_PMD=n
> > CONFIG_RTE_LIBRTE_I40E_PMD=n
> > CONFIG_RTE_LIBRTE_VIRTIO_PMD=y
The Ark/PMD is not supported on those architectures at this time. This is
also reflected in doc/guides/nic/ark.rst
> Can you also please document the device arguments in this file?
> Device Parameters
> -----------------
> "Pkt_gen"
> "Pkt_chkr"
> "Pkt_dir"
doc/guides/nic/ark.rst has been update to include documentation for these
arguments.
> > diff --git a/doc/guides/nics/features/ark.ini
b/doc/guides/nics/features/ark.ini
> ...
> Features can be added with the patch that adds functionality. I believe
> above features not supported with current patch.
The ark.ini file has been removed in this patch and will be added in a
later patch
> > +#
> > +SRCS-y += ark_ethdev.c
> Please use SRCS-y only in comment, for actual usage please prefer:
> SRCS-$(CONFIG_RTE_LIBRTE_ARK_PMD)
Changed per request.
>
> > +#define ARK_TRACE_ON(fmt, ...) \
> > + PMD_DRV_LOG(ERR, fmt, ##__VA_ARGS__)
> > +
> > +#define ARK_TRACE_OFF(fmt, ...) \
> > + do {if (0) PMD_DRV_LOG(ERR, fmt, ##__VA_ARGS__); } while (0)
> why not just "do { } while(0)" ?
A do while body always executes at least once. The if (0) is required.
> This is debug option but prints only "ERR" level log, shouldn't this be
> DEBUG.
> Also there are dpdk wide log level option, helps optimizing out some
> code, if you use only ERR type, you won't be benefiting from it.
Changed to DEBUG
>
> __rte_unused not required.
Removed.
> > + ARK_DEBUG_TRACE("eth_ark_dev_init(struct rte_eth_dev *dev)\n");
> > + gark[0] = ark;
> Is it OK to have this hardcoded index "0"? When there are two instance
> of this device, this value be overwritten by second one.
Code refactored moving this variable from global scope to instance specific.
> dev->data->dev_flags |= RTE_ETH_DEV_DETACHABLE;
I do not understand your comment.
>
> memset not required, since already done by ethdev abstraction layer,
> specially desc_lim values already overwritten below.
Deleted.
>
> > + dev_info->max_rx_pktlen = (16 * 1024) - 128;
> > + dev_info->min_rx_bufsize = 1024;
> Using some macros instead of hardcoded values helps to understand values.
Done
>
>
> The general usage of DEBUG_TRACE is providing backtrace log, function
> enterance / exit informations. I guess, that is why it has been
> controlled by different config option.
> Here what you need looks like regular debugging functions, PMD_DRV_LOG /
> RTE_LOG variant.
I'm unclear on your request or comment. Are you suggesting that we use the
dpdk versions of debug? The advantage of a local version is that they can
be enabled without all the debug traces.
> Can this overflow args variable? Any way to prevent possible crash?
> What happens if somebody, by accident, provides a file as device
> argument which is larger than 256 bytes?
Code has been fixed to avoid overflow.
>> > +/*
>>
>> > + * Although Arkville is a physical device we take advantage of the
virtual
>>
>> > + * device initialization as a per test runtime initialization for
>>
>> > + * regression testing. Parameters are passed into the virtual device
to
>>
>> > + * configure the packet generator, packet director and packet checker.
>>
>> > + */
>>
>> Why not providing these arguments via physical device?
Code has been changed to use dev arg instead of vdev args.
>> +++ b/drivers/net/ark/rte_pmd_ark_version.map
> s/DPDK_2.0/DPDK_17.05
Fixed.
next prev parent reply other threads:[~2017-03-23 19:46 UTC|newest]
Thread overview: 65+ messages / expand[flat|nested] mbox.gz Atom feed top
2017-03-23 1:03 Ed Czeck
2017-03-23 1:03 ` [dpdk-dev] [PATCH v4 2/7] net/ark: HW API part 1 of 3 Ed Czeck
2017-03-23 11:38 ` Ferruh Yigit
2017-03-23 20:33 ` Ed Czeck
2017-03-29 1:05 ` [dpdk-dev] [PATCH v6 2/7] net/ark: Provide API for hardware modules mpu, rqp, and pktdir Ed Czeck
2017-03-29 21:32 ` [dpdk-dev] [PATCH v7 2/7] net/ark: provide API for hardware modules mpu rqp " Ed Czeck
2017-04-04 19:50 ` [dpdk-dev] [PATCH v8 " Ed Czeck
2017-03-23 1:03 ` [dpdk-dev] [PATCH v4 3/7] net/ark: HW API part 2 of 3 Ed Czeck
2017-03-29 1:05 ` [dpdk-dev] [PATCH v6 3/7] net/ark: Provide API for hardware modules udm and ddm Ed Czeck
2017-03-29 21:33 ` [dpdk-dev] [PATCH v7 3/7] net/ark: provide " Ed Czeck
2017-04-04 19:50 ` [dpdk-dev] [PATCH v8 " Ed Czeck
2017-03-23 1:03 ` [dpdk-dev] [PATCH v4 4/7] net/ark: HW API part 3 of 3 Ed Czeck
2017-03-29 1:06 ` [dpdk-dev] [PATCH v6 4/7] net/ark: Provide API for hardware modules pktchkr and pktgen Ed Czeck
2017-03-29 21:34 ` [dpdk-dev] [PATCH v7 4/7] net/ark: provide " Ed Czeck
2017-04-04 19:50 ` [dpdk-dev] [PATCH v8 " Ed Czeck
2017-03-23 1:03 ` [dpdk-dev] [PATCH v4 5/7] net/ark: Packet TX support initial version Ed Czeck
2017-03-23 12:14 ` Ferruh Yigit
2017-03-23 21:44 ` Ed Czeck
2017-03-29 1:06 ` [dpdk-dev] [PATCH v6 " Ed Czeck
2017-03-29 21:34 ` [dpdk-dev] [PATCH v7 5/7] net/ark: packet Tx " Ed Czeck
2017-04-04 19:51 ` [dpdk-dev] [PATCH v8 " Ed Czeck
2017-03-23 1:03 ` [dpdk-dev] [PATCH v4 6/7] net/ark: Packet RX " Ed Czeck
2017-03-23 12:14 ` Ferruh Yigit
2017-03-23 21:51 ` Ed Czeck
2017-03-29 1:06 ` [dpdk-dev] [PATCH v6 " Ed Czeck
2017-03-29 21:35 ` [dpdk-dev] [PATCH v7 6/7] net/ark: packet Rx " Ed Czeck
2017-04-04 19:51 ` [dpdk-dev] [PATCH v8 " Ed Czeck
2017-03-23 1:03 ` [dpdk-dev] [PATCH v4 7/7] net/ark: Arkville PMD component integration Ed Czeck
2017-03-23 12:13 ` Ferruh Yigit
2017-03-23 22:19 ` Ed Czeck
2017-03-28 12:34 ` Ferruh Yigit
2017-03-29 1:07 ` [dpdk-dev] [PATCH v6 " Ed Czeck
2017-03-29 9:54 ` Ferruh Yigit
2017-03-29 21:35 ` [dpdk-dev] [PATCH v7 7/7] net/ark: arkville " Ed Czeck
2017-04-04 19:51 ` [dpdk-dev] [PATCH v8 " Ed Czeck
2017-03-23 11:36 ` [dpdk-dev] [PATCH v4 1/7] net/ark: PMD for Atomic Rules Arkville driver stub Ferruh Yigit
2017-03-23 13:08 ` Ferruh Yigit
2017-03-23 19:46 ` Ed Czeck [this message]
2017-03-28 12:58 ` Ferruh Yigit
2017-03-28 21:11 ` Ed Czeck
2017-03-29 9:42 ` Ferruh Yigit
2017-03-23 22:59 ` [dpdk-dev] [PATCH v5 " Ed Czeck
2017-03-23 23:00 ` [dpdk-dev] [PATCH v5 2/7] net/ark: provide api to hardware module mpu, rqp, and pktdir Ed Czeck
2017-03-28 14:35 ` Ferruh Yigit
2017-03-28 20:14 ` Ed Czeck
2017-03-23 23:00 ` [dpdk-dev] [PATCH v5 3/7] net/ark: provide API hardware module udm and ddm Ed Czeck
2017-03-23 23:00 ` [dpdk-dev] [PATCH v5 4/7] net/ark: prrovide api for hardware module pktchkr and pktgen Ed Czeck
2017-03-23 23:00 ` [dpdk-dev] [PATCH v5 5/7] net/ark: Packet TX support initial version Ed Czeck
2017-03-23 23:01 ` [dpdk-dev] [PATCH v5 6/7] net/ark: Packet RX " Ed Czeck
2017-03-28 14:36 ` Ferruh Yigit
2017-03-28 21:59 ` Ed Czeck
2017-03-23 23:01 ` [dpdk-dev] [PATCH v5 7/7] net/ark: Arkville PMD component integration Ed Czeck
2017-03-28 14:38 ` Ferruh Yigit
2017-03-28 15:00 ` Adrien Mazarguil
2017-03-28 22:42 ` Ed Czeck
2017-03-28 14:34 ` [dpdk-dev] [PATCH v5 1/7] net/ark: PMD for Atomic Rules Arkville driver stub Ferruh Yigit
2017-03-28 22:38 ` Ed Czeck
2017-03-28 14:41 ` Ferruh Yigit
2017-03-29 1:04 ` [dpdk-dev] [PATCH v6 " Ed Czeck
2017-03-29 21:32 ` [dpdk-dev] [PATCH v7 1/7] net/ark: stub PMD for Atomic Rules Arkville Ed Czeck
2017-03-31 14:51 ` Ferruh Yigit
2017-03-31 15:09 ` Shepard Siegel
2017-04-04 20:58 ` Ed Czeck
2017-04-04 19:50 ` [dpdk-dev] [PATCH v8 " Ed Czeck
2017-04-07 10:54 ` Ferruh Yigit
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=CALZ3GuhCCVPFJ9kqxgnecQTNcf9M9uqVxMN+mOJR5VKckScZwA@mail.gmail.com \
--to=ed.czeck@atomicrules.com \
--cc=dev@dpdk.org \
--cc=ferruh.yigit@intel.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
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).