From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: Jakub Kicinski <kuba@kernel.org>
Cc: Joel Stanley <joel@jms.id.au>,
Dylan Hung <dylan_hung@aspeedtech.com>,
"David S . Miller" <davem@davemloft.net>,
netdev@vger.kernel.org,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Po-Yu Chuang <ratbert@faraday-tech.com>,
linux-aspeed <linux-aspeed@lists.ozlabs.org>,
OpenBMC Maillist <openbmc@lists.ozlabs.org>,
BMC-SW <BMC-SW@aspeedtech.com>, Arnd Bergmann <arnd@arndb.de>,
linux-arch@vger.kernel.org
Subject: Re: [PATCH] net: ftgmac100: Fix missing TX-poll issue
Date: Tue, 20 Oct 2020 17:15:42 +1100 [thread overview]
Message-ID: <3ebaa814fe21eb7b4b25a2c9455a34434e0207d6.camel@kernel.crashing.org> (raw)
In-Reply-To: <20201019195723.41a5591f@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
On Mon, 2020-10-19 at 19:57 -0700, Jakub Kicinski wrote:
> > I suspect the problem is that the HW (and yes this would be a HW bug)
> > doesn't order the CPU -> memory and the CPU -> MMIO path.
> >
> > What I think happens is that the store to txde0 is potentially still in
> > a buffer somewhere on its way to memory, gets bypassed by the store to
> > MMIO, causing the MAC to try to read the descriptor, and getting the
> > "old" data from memory.
>
> I see, but in general this sort of a problem should be resolved by
> adding an appropriate memory barrier. And in fact such barrier should
> (these days) be implied by a writel (I'm not 100% clear on why this
> driver uses iowrite, and if it matters).
No, a barrier won't solve this I think.
This is a coherency problem at the fabric/interconnect level. I has to
do with the way they implemented the DMA path from memory to the
ethernet controller using a different "port" of the memory controller
than the one used by the CPU, separately from the MMIO path, with no
proper ordering between those busses. Old school design .... and
broken.
By doing a read back, they probably force the previous write to memory
to get past the point where it will be visible to a subsequent DMA read
by the ethernet controller.
> > It's ... fishy, but that isn't the first time an Aspeed chip has that
> > type of bug (there's a similar one in the USB device controler iirc).
Cheers,
Ben.
WARNING: multiple messages have this Message-ID (diff)
From: Benjamin Herrenschmidt <benh@kernel.crashing.org>
To: Jakub Kicinski <kuba@kernel.org>
Cc: linux-arch@vger.kernel.org, BMC-SW <BMC-SW@aspeedtech.com>,
linux-aspeed <linux-aspeed@lists.ozlabs.org>,
Po-Yu Chuang <ratbert@faraday-tech.com>,
netdev@vger.kernel.org,
OpenBMC Maillist <openbmc@lists.ozlabs.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
Arnd Bergmann <arnd@arndb.de>,
Dylan Hung <dylan_hung@aspeedtech.com>,
"David S . Miller" <davem@davemloft.net>
Subject: Re: [PATCH] net: ftgmac100: Fix missing TX-poll issue
Date: Tue, 20 Oct 2020 17:15:42 +1100 [thread overview]
Message-ID: <3ebaa814fe21eb7b4b25a2c9455a34434e0207d6.camel@kernel.crashing.org> (raw)
In-Reply-To: <20201019195723.41a5591f@kicinski-fedora-pc1c0hjn.dhcp.thefacebook.com>
On Mon, 2020-10-19 at 19:57 -0700, Jakub Kicinski wrote:
> > I suspect the problem is that the HW (and yes this would be a HW bug)
> > doesn't order the CPU -> memory and the CPU -> MMIO path.
> >
> > What I think happens is that the store to txde0 is potentially still in
> > a buffer somewhere on its way to memory, gets bypassed by the store to
> > MMIO, causing the MAC to try to read the descriptor, and getting the
> > "old" data from memory.
>
> I see, but in general this sort of a problem should be resolved by
> adding an appropriate memory barrier. And in fact such barrier should
> (these days) be implied by a writel (I'm not 100% clear on why this
> driver uses iowrite, and if it matters).
No, a barrier won't solve this I think.
This is a coherency problem at the fabric/interconnect level. I has to
do with the way they implemented the DMA path from memory to the
ethernet controller using a different "port" of the memory controller
than the one used by the CPU, separately from the MMIO path, with no
proper ordering between those busses. Old school design .... and
broken.
By doing a read back, they probably force the previous write to memory
to get past the point where it will be visible to a subsequent DMA read
by the ethernet controller.
> > It's ... fishy, but that isn't the first time an Aspeed chip has that
> > type of bug (there's a similar one in the USB device controler iirc).
Cheers,
Ben.
next prev parent reply other threads:[~2020-10-20 6:16 UTC|newest]
Thread overview: 37+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-10-19 7:39 [PATCH] net: ftgmac100: Fix missing TX-poll issue Dylan Hung
2020-10-19 7:39 ` Dylan Hung
2020-10-19 8:57 ` Joel Stanley
2020-10-19 8:57 ` Joel Stanley
2020-10-19 9:19 ` Dylan Hung
2020-10-19 9:19 ` Dylan Hung
2020-10-19 19:00 ` Jakub Kicinski
2020-10-19 19:00 ` Jakub Kicinski
2020-10-19 23:23 ` Benjamin Herrenschmidt
2020-10-19 23:23 ` Benjamin Herrenschmidt
2020-10-20 2:57 ` Jakub Kicinski
2020-10-20 2:57 ` Jakub Kicinski
2020-10-20 6:15 ` Benjamin Herrenschmidt [this message]
2020-10-20 6:15 ` Benjamin Herrenschmidt
2020-10-20 17:24 ` Jakub Kicinski
2020-10-20 17:24 ` Jakub Kicinski
2020-10-20 6:14 ` Dylan Hung
2020-10-20 6:14 ` Dylan Hung
2020-10-20 13:15 ` David Laight
2020-10-20 13:15 ` David Laight
2020-10-20 22:05 ` Benjamin Herrenschmidt
2020-10-20 22:05 ` Benjamin Herrenschmidt
2020-10-20 19:49 ` Arnd Bergmann
2020-10-20 22:10 ` Benjamin Herrenschmidt
2020-10-20 22:10 ` Benjamin Herrenschmidt
2020-10-20 22:25 ` Andrew Jeffery
2020-10-23 13:08 ` Dylan Hung
2020-10-26 22:21 ` Benjamin Herrenschmidt
2020-10-27 2:18 ` Joel Stanley
2020-10-27 2:18 ` Joel Stanley
2020-10-21 7:16 ` Arnd Bergmann
2020-10-21 12:11 ` Arnd Bergmann
2020-10-21 12:11 ` Arnd Bergmann
2020-10-22 7:40 ` Benjamin Herrenschmidt
2020-10-22 7:40 ` Benjamin Herrenschmidt
2020-10-23 8:39 ` Arnd Bergmann
2020-10-23 8:39 ` 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=3ebaa814fe21eb7b4b25a2c9455a34434e0207d6.camel@kernel.crashing.org \
--to=benh@kernel.crashing.org \
--cc=BMC-SW@aspeedtech.com \
--cc=arnd@arndb.de \
--cc=davem@davemloft.net \
--cc=dylan_hung@aspeedtech.com \
--cc=joel@jms.id.au \
--cc=kuba@kernel.org \
--cc=linux-arch@vger.kernel.org \
--cc=linux-aspeed@lists.ozlabs.org \
--cc=linux-kernel@vger.kernel.org \
--cc=netdev@vger.kernel.org \
--cc=openbmc@lists.ozlabs.org \
--cc=ratbert@faraday-tech.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 an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.