From: Jerin Jacob <jerinjacobk@gmail.com>
To: fengchengwen <fengchengwen@huawei.com>
Cc: "Bruce Richardson" <bruce.richardson@intel.com>,
"Morten Brørup" <mb@smartsharesystems.com>,
"Thomas Monjalon" <thomas@monjalon.net>,
"Ferruh Yigit" <ferruh.yigit@intel.com>, dpdk-dev <dev@dpdk.org>,
"Nipun Gupta" <nipun.gupta@nxp.com>,
"Hemant Agrawal" <hemant.agrawal@nxp.com>,
"Maxime Coquelin" <maxime.coquelin@redhat.com>,
"Honnappa Nagarahalli" <honnappa.nagarahalli@arm.com>,
"Jerin Jacob" <jerinj@marvell.com>,
"David Marchand" <david.marchand@redhat.com>,
"Satananda Burla" <sburla@marvell.com>,
"Prasun Kapoor" <pkapoor@marvell.com>
Subject: Re: [dpdk-dev] [RFC PATCH] dmadev: introduce DMA device library
Date: Wed, 23 Jun 2021 12:51:07 +0530 [thread overview]
Message-ID: <CALBAE1OM-LKm1B_z=C==LBXybsH1-v5=JQXzbKB4E=Mu2ka0hQ@mail.gmail.com> (raw)
In-Reply-To: <aceadea0-a786-a666-cf70-4e501d501ba1@huawei.com>
On Wed, Jun 23, 2021 at 9:00 AM fengchengwen <fengchengwen@huawei.com> wrote:
>
> >>>
> >>>>
> >>>>> The above will give better performance and is the best trade-off c
> >>>>> between performance and per transfer variables.
> >>>>
> >>>> We may need to have different APIs for context-aware and context-unaware
> >>>> processing, with which to use determined by the capabilities discovery.
> >>>> Given that for these DMA devices the offload cost is critical, more so than
> >>>> any other dev class I've looked at before, I'd like to avoid having APIs
> >>>> with extra parameters than need to be passed about since that just adds
> >>>> extra CPU cycles to the offload.
> >>>
> >>> If driver does not support additional attributes and/or the
> >>> application does not need it, rte_dmadev_desc_t can be NULL.
> >>> So that it won't have any cost in the datapath. I think, we can go to
> >>> different API
> >>> cases if we can not abstract problems without performance impact.
> >>> Otherwise, it will be too much
> >>> pain for applications.
> >>
> >> Yes, currently we plan to use different API for different case, e.g.
> >> rte_dmadev_memcpy() -- deal with local to local memcopy
> >> rte_dmadev_memset() -- deal with fill with local memory with pattern
> >> maybe:
> >> rte_dmadev_imm_data() --deal with copy very little data
> >> rte_dmadev_p2pcopy() --deal with peer-to-peer copy of diffenet PCIE addr
> >>
> >> These API capabilities will be reflected in the device capability set so that
> >> application could know by standard API.
> >
> >
> > There will be a lot of combination of that it will be like M x N cross
> > base case, It won't scale.
>
> Currently, it is hard to define generic dma descriptor, I think the well-defined
> APIs is feasible.
I would like to understand why not feasible? if we move the
preparation to the slow path.
i.e
struct rte_dmadev_desc defines all the "attributes" of all DMA devices available
using capability. I believe with the scheme, we can scale and
incorporate all features of
all DMA HW without any performance impact.
something like:
struct rte_dmadev_desc {
/* Attributes all DMA transfer available for all HW under capability. */
channel or port;
ops ; // copy, fill etc..
/* impemention opqueue memory as zero length array,
rte_dmadev_desc_prep() update this memory with HW specific information
*/
uint8_t impl_opq[];
}
// allocate the memory for dma decriptor
struct rte_dmadev_desc *rte_dmadev_desc_alloc(devid);
// Convert DPDK specific descriptors to HW specific descriptors in slowpath */
rte_dmadev_desc_prep(devid, struct rte_dmadev_desc *desc);
// Free dma descriptor memory
rte_dmadev_desc_free(devid, struct rte_dmadev_desc *desc )
The above calls in slow path.
Only below call in fastpath.
// Here desc can be NULL(in case you don't need any specific attribute
attached to transfer, if needed, it can be an object which is gone
through rte_dmadev_desc_prep())
rte_dmadev_enq(devid, struct rte_dmadev_desc *desc, void *src, void
*dest, unsigned int len, cookie)
>
> >
> >>
> >>>
> >>> Just to understand, I think, we need to HW capabilities and how to
> >>> have a common API.
> >>> I assume HW will have some HW JOB descriptors which will be filled in
> >>> SW and submitted to HW.
> >>> In our HW, Job descriptor has the following main elements
> >>>
> >>> - Channel // We don't expect the application to change per transfer
> >>> - Source address - It can be scatter-gather too - Will be changed per transfer
> >>> - Destination address - It can be scatter-gather too - Will be changed
> >>> per transfer
> >>> - Transfer Length - - It can be scatter-gather too - Will be changed
> >>> per transfer
> >>> - IOVA address where HW post Job completion status PER Job descriptor
> >>> - Will be changed per transfer
> >>> - Another sideband information related to channel // We don't expect
> >>> the application to change per transfer
> >>> - As an option, Job completion can be posted as an event to
> >>> rte_event_queue too // We don't expect the application to change per
> >>> transfer
> >>
> >> The 'option' field looks like a software interface field, but not HW descriptor.
> >
> > It is in HW descriptor.
>
> The HW is interesting, something like: DMA could send completion direct to EventHWQueue,
> the DMA and EventHWQueue are link in the hardware range, rather than by software.
Yes.
>
> Could you provide public driver of this HW ? So we could know more about it's working
> mechanism and software-hardware collaboration.
http://code.dpdk.org/dpdk/v21.05/source/drivers/raw/octeontx2_dma/otx2_dpi_rawdev.h#L149
is the DMA instruction header.
next prev parent reply other threads:[~2021-06-23 7:21 UTC|newest]
Thread overview: 79+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-15 13:22 Chengwen Feng
2021-06-15 16:38 ` Bruce Richardson
2021-06-16 7:09 ` Morten Brørup
2021-06-16 10:17 ` fengchengwen
2021-06-16 12:09 ` Morten Brørup
2021-06-16 13:06 ` Bruce Richardson
2021-06-16 14:37 ` Jerin Jacob
2021-06-17 9:15 ` Bruce Richardson
2021-06-18 5:52 ` Jerin Jacob
2021-06-18 9:41 ` fengchengwen
2021-06-22 17:25 ` Jerin Jacob
2021-06-23 3:30 ` fengchengwen
2021-06-23 7:21 ` Jerin Jacob [this message]
2021-06-23 9:37 ` Bruce Richardson
2021-06-23 11:40 ` Jerin Jacob
2021-06-23 14:19 ` Bruce Richardson
2021-06-24 6:49 ` Jerin Jacob
2021-06-23 9:41 ` Bruce Richardson
2021-06-23 10:10 ` Morten Brørup
2021-06-23 11:46 ` Jerin Jacob
2021-06-23 14:22 ` Bruce Richardson
2021-06-18 9:55 ` Bruce Richardson
2021-06-22 17:31 ` Jerin Jacob
2021-06-22 19:17 ` Bruce Richardson
2021-06-23 7:00 ` Jerin Jacob
2021-06-16 9:41 ` fengchengwen
2021-06-16 17:31 ` Bruce Richardson
2021-06-16 18:08 ` Jerin Jacob
2021-06-16 19:13 ` Bruce Richardson
2021-06-17 7:42 ` Jerin Jacob
2021-06-17 8:00 ` Bruce Richardson
2021-06-18 5:16 ` Jerin Jacob
2021-06-18 10:03 ` Bruce Richardson
2021-06-22 17:36 ` Jerin Jacob
2021-06-17 9:48 ` fengchengwen
2021-06-17 11:02 ` Bruce Richardson
2021-06-17 14:18 ` Bruce Richardson
2021-06-18 8:52 ` fengchengwen
2021-06-18 9:30 ` Bruce Richardson
2021-06-22 17:51 ` Jerin Jacob
2021-06-23 3:50 ` fengchengwen
2021-06-23 11:00 ` Jerin Jacob
2021-06-23 14:56 ` Bruce Richardson
2021-06-24 12:19 ` fengchengwen
2021-06-26 3:59 ` [dpdk-dev] dmadev discussion summary fengchengwen
2021-06-28 10:00 ` Bruce Richardson
2021-06-28 11:14 ` Ananyev, Konstantin
2021-06-28 12:53 ` Bruce Richardson
2021-07-02 13:31 ` fengchengwen
2021-07-01 15:01 ` Jerin Jacob
2021-07-01 16:33 ` Bruce Richardson
2021-07-02 7:39 ` Morten Brørup
2021-07-02 10:05 ` Bruce Richardson
2021-07-02 13:45 ` fengchengwen
2021-07-02 14:57 ` Morten Brørup
2021-07-03 0:32 ` fengchengwen
2021-07-03 8:53 ` Morten Brørup
2021-07-03 9:08 ` Jerin Jacob
2021-07-03 12:24 ` Morten Brørup
2021-07-04 7:43 ` Jerin Jacob
2021-07-05 10:28 ` Morten Brørup
2021-07-06 7:11 ` fengchengwen
2021-07-03 9:45 ` fengchengwen
2021-07-03 12:00 ` Morten Brørup
2021-07-04 7:34 ` Jerin Jacob
2021-07-02 7:07 ` Liang Ma
2021-07-02 13:59 ` fengchengwen
2021-06-24 7:03 ` [dpdk-dev] [RFC PATCH] dmadev: introduce DMA device library Jerin Jacob
2021-06-24 7:59 ` Morten Brørup
2021-06-24 8:05 ` Jerin Jacob
2021-06-23 5:34 ` Hu, Jiayu
2021-06-23 11:07 ` Jerin Jacob
2021-06-16 2:17 ` Wang, Haiyue
2021-06-16 8:04 ` Bruce Richardson
2021-06-16 8:16 ` Wang, Haiyue
2021-06-16 12:14 ` David Marchand
2021-06-16 13:11 ` Bruce Richardson
2021-06-16 16:48 ` Honnappa Nagarahalli
2021-06-16 19:10 ` Bruce Richardson
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='CALBAE1OM-LKm1B_z=C==LBXybsH1-v5=JQXzbKB4E=Mu2ka0hQ@mail.gmail.com' \
--to=jerinjacobk@gmail.com \
--cc=bruce.richardson@intel.com \
--cc=david.marchand@redhat.com \
--cc=dev@dpdk.org \
--cc=fengchengwen@huawei.com \
--cc=ferruh.yigit@intel.com \
--cc=hemant.agrawal@nxp.com \
--cc=honnappa.nagarahalli@arm.com \
--cc=jerinj@marvell.com \
--cc=maxime.coquelin@redhat.com \
--cc=mb@smartsharesystems.com \
--cc=nipun.gupta@nxp.com \
--cc=pkapoor@marvell.com \
--cc=sburla@marvell.com \
--cc=thomas@monjalon.net \
/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).