From: Jan Glauber <jglauber@cavium.com>
To: Wolfram Sang <wsa@the-dreams.de>
Cc: linux-kernel@vger.kernel.org, linux-i2c@vger.kernel.org,
David Daney <ddaney@caviumnetworks.com>,
David Daney <david.daney@cavium.com>,
Jan Glauber <jglauber@cavium.com>
Subject: [PATCH v3 11/14] i2c-octeon: Add workaround for broken irqs on CN3860
Date: Mon, 7 Mar 2016 16:10:54 +0100 [thread overview]
Message-ID: <20fb6eb5fbc0384f51a7e2972d903c981a65f545.1457362545.git.jglauber@cavium.com> (raw)
In-Reply-To: <cover.1457362545.git.jglauber@cavium.com>
In-Reply-To: <cover.1457362545.git.jglauber@cavium.com>
From: David Daney <david.daney@cavium.com>
CN3860 does not interrupt the CPU when the i2c status changes. If
we get a timeout, and see the status has in fact changed, we know we
have this problem, and drop back to polling.
Signed-off-by: David Daney <ddaney@caviumnetworks.com>
Signed-off-by: Jan Glauber <jglauber@cavium.com>
---
drivers/i2c/busses/i2c-octeon.c | 46 +++++++++++++++++++++++++++++++++++++++++
1 file changed, 46 insertions(+)
diff --git a/drivers/i2c/busses/i2c-octeon.c b/drivers/i2c/busses/i2c-octeon.c
index afca316..6868802 100644
--- a/drivers/i2c/busses/i2c-octeon.c
+++ b/drivers/i2c/busses/i2c-octeon.c
@@ -112,6 +112,7 @@ struct octeon_i2c {
void __iomem *twsi_base;
struct device *dev;
bool hlc_enabled;
+ bool broken_irq_mode;
void (*int_en)(struct octeon_i2c *);
void (*int_dis)(struct octeon_i2c *);
void (*hlc_int_en)(struct octeon_i2c *);
@@ -378,10 +379,33 @@ static int octeon_i2c_wait(struct octeon_i2c *i2c)
long time_left;
int first = 1;
+ if (i2c->broken_irq_mode) {
+ /*
+ * Some chip revisions seem to not assert the irq in
+ * the interrupt controller. So we must poll for the
+ * IFLG change.
+ */
+ u64 end = get_jiffies_64() + i2c->adap.timeout;
+
+ while (!octeon_i2c_test_iflg(i2c) &&
+ time_before64(get_jiffies_64(), end))
+ udelay(50);
+
+ return octeon_i2c_test_iflg(i2c) ? 0 : -ETIMEDOUT;
+ }
+
i2c->int_en(i2c);
time_left = wait_event_timeout(i2c->queue, poll_iflg(i2c, &first),
i2c->adap.timeout);
i2c->int_dis(i2c);
+
+ if (time_left <= 0 && OCTEON_IS_MODEL(OCTEON_CN38XX) &&
+ octeon_i2c_test_iflg(i2c)) {
+ dev_err(i2c->dev,
+ "broken irq connection detected, switching to polling mode.\n");
+ i2c->broken_irq_mode = true;
+ return 0;
+ }
if (!time_left) {
dev_dbg(i2c->dev, "%s: timeout\n", __func__);
return -ETIMEDOUT;
@@ -516,6 +540,21 @@ static int octeon_i2c_hlc_wait(struct octeon_i2c *i2c)
{
int result;
+ if (i2c->broken_irq_mode) {
+ /*
+ * Some cn38xx boards did not assert the irq in
+ * the interrupt controller. So we must poll for the
+ * IFLG change.
+ */
+ u64 end = get_jiffies_64() + i2c->adap.timeout;
+
+ while (!octeon_i2c_hlc_test_ready(i2c) &&
+ time_before64(get_jiffies_64(), end))
+ udelay(50);
+
+ return octeon_i2c_hlc_test_ready(i2c) ? 0 : -ETIMEDOUT;
+ }
+
i2c->hlc_int_en(i2c);
result = wait_event_interruptible_timeout(i2c->queue,
octeon_i2c_hlc_test_ready(i2c),
@@ -523,6 +562,13 @@ static int octeon_i2c_hlc_wait(struct octeon_i2c *i2c)
i2c->hlc_int_dis(i2c);
if (!result)
octeon_i2c_hlc_int_clear(i2c);
+ if (result <= 0 && OCTEON_IS_MODEL(OCTEON_CN38XX) &&
+ octeon_i2c_hlc_test_ready(i2c)) {
+ dev_err(i2c->dev, "broken irq connection detected, switching to polling mode.\n");
+ i2c->broken_irq_mode = true;
+ return 0;
+ }
+
if (result < 0) {
dev_dbg(i2c->dev, "%s: wait interrupted\n", __func__);
return result;
--
1.9.1
next prev parent reply other threads:[~2016-03-07 15:12 UTC|newest]
Thread overview: 24+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-03-07 15:10 [PATCH v3 00/14] i2c-octeon and i2c-thunderx drivers Jan Glauber
2016-03-07 15:10 ` [PATCH v3 01/14] i2c-octeon: Cleanup kerneldoc comments Jan Glauber
2016-03-12 15:35 ` Wolfram Sang
2016-03-07 15:10 ` [PATCH v3 02/14] i2c-octeon: Cleanup i2c-octeon driver Jan Glauber
2016-03-12 15:35 ` Wolfram Sang
2016-03-14 8:52 ` Jan Glauber
2016-03-07 15:10 ` [PATCH v3 03/14] i2c-octeon: Cleanup resource allocation code Jan Glauber
2016-03-12 15:37 ` Wolfram Sang
2016-03-14 9:21 ` Jan Glauber
2016-03-07 15:10 ` [PATCH v3 04/14] i2c-octeon: Support I2C_M_RECV_LEN Jan Glauber
2016-03-12 15:44 ` Wolfram Sang
2016-03-07 15:10 ` [PATCH v3 05/14] i2c-octeon: Make adapter timeout tunable Jan Glauber
2016-03-12 15:46 ` Wolfram Sang
2016-03-14 12:45 ` Jan Glauber
2016-03-14 17:21 ` Swain, Peter
2016-03-07 15:10 ` [PATCH v3 06/14] i2c-octeon: Enable high-level controller and improve on bus contention Jan Glauber
2016-03-07 15:10 ` [PATCH v3 07/14] dt-bindings: i2c: Add Octeon cn78xx TWSI Jan Glauber
2016-03-07 15:10 ` [PATCH v3 08/14] i2c-octeon: Add support for cn78XX chips Jan Glauber
2016-03-07 15:10 ` [PATCH v3 09/14] i2c-octeon: Flush TWSI writes with readback Jan Glauber
2016-03-07 15:10 ` [PATCH v3 10/14] i2c-octeon: Faster operation when IFLG signals late Jan Glauber
2016-03-07 15:10 ` Jan Glauber [this message]
2016-03-07 15:10 ` [PATCH v3 12/14] i2c-octeon: Split the driver into two parts Jan Glauber
2016-03-07 15:10 ` [PATCH v3 13/14] i2c-thunderx: Add i2c driver for ThunderX SOC Jan Glauber
2016-03-07 15:10 ` [PATCH v3 14/14] i2c-thunderx: Add smbus support Jan Glauber
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=20fb6eb5fbc0384f51a7e2972d903c981a65f545.1457362545.git.jglauber@cavium.com \
--to=jglauber@cavium.com \
--cc=david.daney@cavium.com \
--cc=ddaney@caviumnetworks.com \
--cc=linux-i2c@vger.kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=wsa@the-dreams.de \
/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).