From: Andy Haydon <ahaydon@moog.com>
To: linux-can@vger.kernel.org
Subject: Re: c_can driver sometimes sends first two bytes filled with zeros
Date: Fri, 10 Jun 2016 10:49:57 +0000 (UTC) [thread overview]
Message-ID: <loom.20160610T122431-248@post.gmane.org> (raw)
In-Reply-To: 574EDEA3.5090204@grandegger.com
Hi Wolfgang,
Wolfgang Grandegger <wg <at> grandegger.com> writes:
>
> Making "priv->write_regs() atomic is something I would try as well. My
> suspicion is that something goes wrong writing(/reading) the controller
> registers.
We have also been looking at a similar issue over the last few days. In our
case, bytes 0 and 1 and bytes 4 and 5 of an 8 byte message can be corrupted
randomly - not necessarily zeroed. Bytes 2 and 3 and bytes 6 and 7 are never
corrupted. The Tx task is running as a high priority real time thread in
PREEMPT_RT.
I've found that so far the only thing that fixes this issue is to perform
the writes to the data registers as 32 bit operations something like this
(for 3.18.20):
--- a/drivers/net/can/c_can/c_can.c 2015-08-07 20:08:04.000000000 +0100
+++ b/drivers/net/can/c_can/c_can.c 2016-06-09 14:46:51.497895357 +0100
@@ -331,9 +332,19 @@
priv->write_reg(priv, C_CAN_IFACE(MSGCTRL_REG, iface), ctrl);
- for (i = 0; i < frame->can_dlc; i += 2) {
- priv->write_reg(priv, C_CAN_IFACE(DATA1_REG, iface) + i / 2,
- frame->data[i] | (frame->data[i + 1] << 8));
+// this doesn't work for Cyclone V
+// for (i = 0; i < frame->can_dlc; i += 2) {
+// priv->write_reg(priv, C_CAN_IFACE(DATA1_REG, iface) + i / 2,
+// frame->data[i] | (frame->data[i + 1] << 8));
+
+// replaced with this as a workaround
+ u32 data = 0;
+ for (i = 0; i < frame->can_dlc; i += 4) {
+ data = (u32)frame->data[i];
+ data |= (u32)frame->data[i + 1] << 8;
+ data |= (u32)frame->data[i + 2] << 16;
+ data |= (u32)frame->data[i + 3] << 24;
+ priv->write_reg32(priv, C_CAN_IFACE(DATA1_REG, iface) + i /
2, data);
}
}
At the moment I can't explain this behaviour, or if this is really the only
way to fix the issue for the Cyclone V.
The tests that I've done so far seem to indicate that the data is correct
in the driver right up to the point at which it is actually written to the
CAN peripheral registers, so I think that your suspicion is correct.
Regards,
Andy
next prev parent reply other threads:[~2016-06-10 10:55 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-05-12 9:23 c_can driver sometimes sends first two bytes filled with zeros Richard Andrysek
2016-05-16 18:14 ` Thor Thayer
2016-05-17 17:18 ` AW: " Richard Andrysek
2016-05-18 15:35 ` Thor Thayer
[not found] ` <0120733A154AE74CA608A286CE7FFD2621D9CB60@rg-contact.RG.local>
2016-05-19 23:00 ` AW: " Thor Thayer
2016-05-20 12:01 ` AW: " Richard Andrysek
2016-05-23 14:22 ` Thor Thayer
2016-05-23 18:19 ` Wolfgang Grandegger
2016-06-01 9:40 ` AW: " Richard Andrysek
2016-06-01 13:09 ` Wolfgang Grandegger
2016-06-10 10:49 ` Andy Haydon [this message]
2016-06-10 12:55 ` Wolfgang Grandegger
2016-06-10 13:12 ` Andy Haydon
2016-06-10 13:36 ` Wolfgang Grandegger
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=loom.20160610T122431-248@post.gmane.org \
--to=ahaydon@moog.com \
--cc=linux-can@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 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.