From: Slava Ovsiienko <viacheslavo@nvidia.com>
To: Honnappa Nagarahalli <Honnappa.Nagarahalli@arm.com>,
"dev@dpdk.org" <dev@dpdk.org>,
Ruifeng Wang <Ruifeng.Wang@arm.com>,
Matan Azrad <matan@nvidia.com>,
Shahaf Shuler <shahafs@nvidia.com>
Cc: nd <nd@arm.com>, Matan Azrad <matan@nvidia.com>,
"stable@dpdk.org" <stable@dpdk.org>, nd <nd@arm.com>
Subject: RE: [PATCH v2] net/mlx5: use just sufficient barrier for Arm platforms
Date: Tue, 7 Mar 2023 16:07:42 +0000 [thread overview]
Message-ID: <DM6PR12MB3753701CC9692D1FB31D2FE7DFB79@DM6PR12MB3753.namprd12.prod.outlook.com> (raw)
In-Reply-To: <DBAPR08MB5814B28C6BADA10F455E01D898049@DBAPR08MB5814.eurprd08.prod.outlook.com>
[-- Attachment #1: Type: text/plain, Size: 4712 bytes --]
Hi, Honnappa
I'm sorry for delay - I had to play with patch, to compile this one with assembly listings
and check what code is actually generated on x86/ARM.
On x86 there is no difference at all (with and w/o patch), so no objection from my side.
On ARM we have:
w/o patch: dmb oshld
with patch: dmb ishld
What is the purpose of the barrier - to not allow the CQE read access reordering.
On x86, "compiler barrier" is quite enough (due to strong load/store ordering).
On ARM load/store might be reordered, AFAIU.
CQE resides in the host memory and can be directly written by the NIC via PCIe.
(In my understanding it can cause core cache(s) invalidations).
I have a question - should we consider this as outer sharable domain?
Is it safe to have a barrier for inner domain only in our case?
We have updated the cqe_check() routine, sorry for this. Could you, please,
update the patch and send v3?
With best regards,
Slava
> -----Original Message-----
> From: Honnappa Nagarahalli <Honnappa.Nagarahalli@arm.com>
> Sent: вторник, 15 ноября 2022 г. 03:46
> To: Slava Ovsiienko <viacheslavo@nvidia.com>; dev@dpdk.org; Ruifeng Wang
> <Ruifeng.Wang@arm.com>; Matan Azrad <matan@nvidia.com>; Shahaf Shuler
> <shahafs@nvidia.com>
> Cc: nd <nd@arm.com>; Matan Azrad <matan@nvidia.com>; stable@dpdk.org;
> nd <nd@arm.com>
> Subject: RE: [PATCH v2] net/mlx5: use just sufficient barrier for Arm platforms
>
> <snip>
>
> >
> > >
> > > Hi, Honnappa
> > Hi Slava, thanks for the feedback.
> >
> > >
> > > We discussed the barrier here:
> > > http://patches.dpdk.org/project/dpdk/patch/20210606164948.35997-1-
> > > honnappa.nagarahalli@arm.com/
> > Yes, I have changed the patch according to the discussion. i.e.
> > barrier is needed, but different (inner sharable domain) barrier is required.
> >
> > >
> > > (BTW, it is good practice to keep the reference to previous patch
> > > versions below Commit Message of the next ones).
> > >
> > > This barrier is not about compiler ordering, it is about external HW
> > > agent memory action completions.
> > > So, I'm not sure the rte_atomic_thread_fence() is safe for x86 -
> > > patch impacts
> > > x86 as well.
> > The earlier barrier 'rte_io_rmb()', resolves to a compiler barrier on x86 [1].
> > The rte_atomic_thread_fence(__ATOMIC_ACQUIRE) on x86 also acts as a
> > compiler barrier. So, there is no change for x86.
> >
> >
> > [1]
> > https://github.com/DPDK/dpdk/blob/main/lib/eal/x86/include/rte_atomic.
> > h#
> > L80
> Hi Slava, any more comments on this?
>
> >
> > >
> > > With best regards,
> > > Slava
> > >
> > > > -----Original Message-----
> > > > From: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
> > > > Sent: Tuesday, August 30, 2022 23:01
> > > > To: dev@dpdk.org; honnappa.nagarahalli@arm.com;
> > > ruifeng.wang@arm.com;
> > > > Matan Azrad <matan@nvidia.com>; Shahaf Shuler
> > <shahafs@nvidia.com>;
> > > > Slava Ovsiienko <viacheslavo@nvidia.com>
> > > > Cc: nd@arm.com; Matan Azrad <matan@nvidia.com>; stable@dpdk.org
> > > > Subject: [PATCH v2] net/mlx5: use just sufficient barrier for Arm
> > > > platforms
> > > >
> > > > cqe->op_own indicates if the CQE is owned by the NIC. The rest of
> > > > the fields in CQE should be read only after op_own is read. On Arm
> > > > platforms using "dmb ishld" is sufficient to enforce this.
> > > >
> > > > Fixes: 88c0733535d6 ("net/mlx5: extend Rx completion with error
> > > > handling")
> > > > Cc: matan@mellanox.com
> > > > Cc: stable@dpdk.org
> > > >
> > > > Signed-off-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
> > > > Reviewed-by: Ruifeng Wang <ruifeng.wang@arm.com>
> > > > ---
> > > > drivers/common/mlx5/mlx5_common.h | 6 +++++-
> > > > 1 file changed, 5 insertions(+), 1 deletion(-)
> > > >
> > > > diff --git a/drivers/common/mlx5/mlx5_common.h
> > > > b/drivers/common/mlx5/mlx5_common.h
> > > > index 5028a05b49..ac2e85b15f 100644
> > > > --- a/drivers/common/mlx5/mlx5_common.h
> > > > +++ b/drivers/common/mlx5/mlx5_common.h
> > > > @@ -195,7 +195,11 @@ check_cqe(volatile struct mlx5_cqe *cqe,
> > > > const uint16_t cqes_n,
> > > >
> > > > if (unlikely((op_owner != (!!(idx))) || (op_code ==
> > > > MLX5_CQE_INVALID)))
> > > > return MLX5_CQE_STATUS_HW_OWN;
> > > > - rte_io_rmb();
> > > > + /* Prevent speculative reading of other fields in CQE until
> > > > + * CQE is valid.
> > > > + */
> > > > + rte_atomic_thread_fence(__ATOMIC_ACQUIRE);
> > > > +
> > > > if (unlikely(op_code == MLX5_CQE_RESP_ERR ||
> > > > op_code == MLX5_CQE_REQ_ERR))
> > > > return MLX5_CQE_STATUS_ERR;
> > > > --
> > > > 2.17.1
[-- Attachment #2: winmail.dat --]
[-- Type: application/ms-tnef, Size: 17159 bytes --]
next prev parent reply other threads:[~2023-03-07 16:07 UTC|newest]
Thread overview: 13+ messages / expand[flat|nested] mbox.gz Atom feed top
2021-06-06 16:49 [dpdk-stable] [PATCH] net/mlx5: remove unwanted barrier Honnappa Nagarahalli
2021-07-01 12:52 ` Slava Ovsiienko
2021-07-02 21:51 ` Honnappa Nagarahalli
2022-08-30 20:00 ` [PATCH v2] net/mlx5: use just sufficient barrier for Arm platforms Honnappa Nagarahalli
2022-09-27 6:34 ` Slava Ovsiienko
2022-09-27 21:03 ` Honnappa Nagarahalli
2022-11-15 1:45 ` Honnappa Nagarahalli
2023-03-07 16:07 ` Slava Ovsiienko [this message]
2023-03-09 2:42 ` Honnappa Nagarahalli
2022-09-27 21:06 ` Honnappa Nagarahalli
2023-03-09 2:32 ` [PATCH v3] net/mlx5: use just sufficient barrier for ARM platforms Honnappa Nagarahalli
2023-03-09 15:44 ` Slava Ovsiienko
2023-03-19 13:32 ` Raslan Darawsheh
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=DM6PR12MB3753701CC9692D1FB31D2FE7DFB79@DM6PR12MB3753.namprd12.prod.outlook.com \
--to=viacheslavo@nvidia.com \
--cc=Honnappa.Nagarahalli@arm.com \
--cc=Ruifeng.Wang@arm.com \
--cc=dev@dpdk.org \
--cc=matan@nvidia.com \
--cc=nd@arm.com \
--cc=shahafs@nvidia.com \
--cc=stable@dpdk.org \
/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).