From: Joel Stanley <joel@jms.id.au>
To: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Cc: Jakub Kicinski <kuba@kernel.org>,
"David S . Miller" <davem@davemloft.net>,
Dylan Hung <dylan_hung@aspeedtech.com>,
netdev@vger.kernel.org,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
linux-aspeed <linux-aspeed@lists.ozlabs.org>
Subject: Re: [PATCH] net: ftgmac100: Ensure tx descriptor updates are visible
Date: Wed, 21 Oct 2020 03:37:02 +0000 [thread overview]
Message-ID: <CACPK8XfmdUdoke8q=z02ijk89=3ZezTfjOmcr9PX3jnpmvAZPw@mail.gmail.com> (raw)
In-Reply-To: <86480db3977cfbf6750209c34a28c8f042be55fb.camel@kernel.crashing.org>
On Wed, 21 Oct 2020 at 00:00, Benjamin Herrenschmidt
<benh@kernel.crashing.org> wrote:
>
> On Wed, 2020-10-21 at 08:36 +1030, Joel Stanley wrote:
> > We must ensure the tx descriptor updates are visible before updating
> > the tx pointer.
> >
> > This resolves the tx hangs observed on the 2600 when running iperf:
>
> To clarify the comment here. This doesn't ensure they are visible to
> the hardware but to other CPUs. This is the ordering vs start_xmit and
> tx_complete.
Thanks. Let me know if this makes sense, or if I'm completely off the mark.
How is this for the commit message:
This resolves the tx hangs observed on the 2600 when running iperf.
This is ensuring the setting of the OWN bit in txdes0 of the
descriptor is visible to other CPUs before updating the pointer. Doing
this provides ordering between start_xmit and tx_complete.
and then I'll put:
/* Ensure the descriptor config is visible to other CPUs before setting
* the tx pointer. This ensures ordering against start_xmit which checks
* the OWN bit before proceeding.
*/
and similar for tx_complete?
>
> Cheers,
> Ben.
>
> > root@ast2600:~# iperf3 -c 192.168.86.146 -R
> > Connecting to host 192.168.86.146, port 5201
> > Reverse mode, remote host 192.168.86.146 is sending
> > [ 5] local 192.168.86.173 port 43886 connected to 192.168.86.146
> > port 5201
> > [ ID] Interval Transfer Bitrate
> > [ 5] 0.00-1.00 sec 90.7 MBytes 760 Mbits/sec
> > [ 5] 1.00-2.00 sec 91.7 MBytes 769 Mbits/sec
> > [ 5] 2.00-3.00 sec 91.7 MBytes 770 Mbits/sec
> > [ 5] 3.00-4.00 sec 91.7 MBytes 769 Mbits/sec
> > [ 5] 4.00-5.00 sec 91.8 MBytes 771 Mbits/sec
> > [ 5] 5.00-6.00 sec 91.8 MBytes 771 Mbits/sec
> > [ 5] 6.00-7.00 sec 91.9 MBytes 771 Mbits/sec
> > [ 5] 7.00-8.00 sec 91.4 MBytes 767 Mbits/sec
> > [ 5] 8.00-9.00 sec 91.3 MBytes 766 Mbits/sec
> > [ 5] 9.00-10.00 sec 91.9 MBytes 771 Mbits/sec
> > [ 5] 10.00-11.00 sec 91.8 MBytes 770 Mbits/sec
> > [ 5] 11.00-12.00 sec 91.8 MBytes 770 Mbits/sec
> > [ 5] 12.00-13.00 sec 90.6 MBytes 761 Mbits/sec
> > [ 5] 13.00-14.00 sec 45.2 KBytes 370 Kbits/sec
> > [ 5] 14.00-15.00 sec 0.00 Bytes 0.00 bits/sec
> > [ 5] 15.00-16.00 sec 0.00 Bytes 0.00 bits/sec
> > [ 5] 16.00-17.00 sec 0.00 Bytes 0.00 bits/sec
> > [ 5] 17.00-18.00 sec 0.00 Bytes 0.00 bits/sec
> > [ 67.031671] ------------[ cut here ]------------
> > [ 67.036870] WARNING: CPU: 1 PID: 0 at net/sched/sch_generic.c:442
> > dev_watchdog+0x2dc/0x300
> > [ 67.046123] NETDEV WATCHDOG: eth2 (ftgmac100): transmit queue 0
> > timed out
> >
> > Fixes: 52c0cae87465 ("ftgmac100: Remove tx descriptor accessors")
> > Signed-off-by: Joel Stanley <joel@jms.id.au>
> > ---
> > drivers/net/ethernet/faraday/ftgmac100.c | 10 ++++++++++
> > 1 file changed, 10 insertions(+)
> >
> > diff --git a/drivers/net/ethernet/faraday/ftgmac100.c
> > b/drivers/net/ethernet/faraday/ftgmac100.c
> > index 331d4bdd4a67..15cdfeb135b0 100644
> > --- a/drivers/net/ethernet/faraday/ftgmac100.c
> > +++ b/drivers/net/ethernet/faraday/ftgmac100.c
> > @@ -653,6 +653,11 @@ static bool ftgmac100_tx_complete_packet(struct
> > ftgmac100 *priv)
> > ftgmac100_free_tx_packet(priv, pointer, skb, txdes, ctl_stat);
> > txdes->txdes0 = cpu_to_le32(ctl_stat & priv-
> > >txdes0_edotr_mask);
> >
> > + /* Ensure the descriptor config is visible before setting the
> > tx
> > + * pointer.
> > + */
> > + smp_wmb();
> > +
> > priv->tx_clean_pointer = ftgmac100_next_tx_pointer(priv,
> > pointer);
> >
> > return true;
> > @@ -806,6 +811,11 @@ static netdev_tx_t
> > ftgmac100_hard_start_xmit(struct sk_buff *skb,
> > dma_wmb();
> > first->txdes0 = cpu_to_le32(f_ctl_stat);
> >
> > + /* Ensure the descriptor config is visible before setting the
> > tx
> > + * pointer.
> > + */
> > + smp_wmb();
> > +
> > /* Update next TX pointer */
> > priv->tx_pointer = pointer;
> >
>
next prev parent reply other threads:[~2020-10-21 3:37 UTC|newest]
Thread overview: 10+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-10-20 22:06 [PATCH] net: ftgmac100: Ensure tx descriptor updates are visible Joel Stanley
2020-10-21 0:00 ` Benjamin Herrenschmidt
2020-10-21 3:37 ` Joel Stanley [this message]
2020-10-21 8:18 ` David Laight
2020-10-22 7:37 ` Benjamin Herrenschmidt
2020-10-21 12:40 ` Arnd Bergmann
2020-10-22 6:35 ` Joel Stanley
2020-10-22 7:41 ` Benjamin Herrenschmidt
2020-10-28 4:47 ` Joel Stanley
2020-10-28 10:51 ` Arnd Bergmann
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='CACPK8XfmdUdoke8q=z02ijk89=3ZezTfjOmcr9PX3jnpmvAZPw@mail.gmail.com' \
--to=joel@jms.id.au \
--cc=benh@kernel.crashing.org \
--cc=davem@davemloft.net \
--cc=dylan_hung@aspeedtech.com \
--cc=kuba@kernel.org \
--cc=linux-aspeed@lists.ozlabs.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.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).