From: Javier Martinez Canillas <javier@osg.samsung.com>
To: Andrzej Hajda <a.hajda@samsung.com>,
Wolfram Sang <wsa@the-dreams.de>,
Krzysztof Kozlowski <krzk@kernel.org>,
linux-i2c@vger.kernel.org, linux-samsung-soc@vger.kernel.org
Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com>,
Marek Szyprowski <m.szyprowski@samsung.com>,
Andi Shyti <andi.shyti@samsung.com>
Subject: Re: [PATCH v2] i2c: exynos5: fix arbitration lost handling
Date: Wed, 8 Mar 2017 17:05:33 -0300 [thread overview]
Message-ID: <2fb11816-85e7-14f6-7b61-5d3eba6096aa@osg.samsung.com> (raw)
In-Reply-To: <1487761474-11547-1-git-send-email-a.hajda@samsung.com>
Hello Andrzej,
On 02/22/2017 08:04 AM, Andrzej Hajda wrote:
> In case of arbitration lost adequate interrupt sometimes is not signaled.
> As a result transfer timeouts and is not retried, as it should. To avoid
> such cases code is added to check transaction status in case of every
> interrupt.
>
> Signed-off-by: Andrzej Hajda <a.hajda@samsung.com>
> ---
This patch causes regressions on Exynos5 boards (at least I noticed it in
Exynos5800 Peach Pi board). I see transmission timeouts when accessing I2C
device registers, i.e:
$ cat /sys/class/rtc/rtc0/time
[ 25.924594] exynos5-hsi2c 12e10000.i2c: rx timeout
[ 65.028365] max77686-rtc max77802-rtc: Fail to read time reg(-22)
cat: /sys/class/rtc/rtc0/time: Invalid argument
Doing a partial revert of $SUBJECT (reading I2C_TRANS_STATUS register when
it was before) "fixes" the problem [0].
I've read the Exynos5800 and Exynos5433 SoC manuals and couldn't figure out
what's causing it. I first thought that the solution was to make I2C core
to retry when the TRANSFER_DONE_AUTO field isn't set [1] but that only makes
the issue less frequent, it still fails. I also tried to increment the retry
count [2] but it also didn't help.
Any ideas about what could be happening here?
[0]:
diff --git a/drivers/i2c/busses/i2c-exynos5.c b/drivers/i2c/busses/i2c-exynos5.c
index cbd93ce0661f..736a82472101 100644
--- a/drivers/i2c/busses/i2c-exynos5.c
+++ b/drivers/i2c/busses/i2c-exynos5.c
@@ -457,7 +457,6 @@ static irqreturn_t exynos5_i2c_irq(int irqno, void *dev_id)
int_status = readl(i2c->regs + HSI2C_INT_STATUS);
writel(int_status, i2c->regs + HSI2C_INT_STATUS);
- trans_status = readl(i2c->regs + HSI2C_TRANS_STATUS);
/* handle interrupt related to the transfer status */
if (i2c->variant->hw == HSI2C_EXYNOS7) {
@@ -482,11 +481,13 @@ static irqreturn_t exynos5_i2c_irq(int irqno, void *dev_id)
goto stop;
}
+ trans_status = readl(i2c->regs + HSI2C_TRANS_STATUS);
if ((trans_status & HSI2C_MASTER_ST_MASK) == HSI2C_MASTER_ST_LOSE) {
i2c->state = -EAGAIN;
goto stop;
}
} else if (int_status & HSI2C_INT_I2C) {
+ trans_status = readl(i2c->regs + HSI2C_TRANS_STATUS);
if (trans_status & HSI2C_NO_DEV_ACK) {
dev_dbg(i2c->dev, "No ACK from device\n");
i2c->state = -ENXIO;
[1]:
>From 3c60d3a0356563b33b739d309a4f64c003d83053 Mon Sep 17 00:00:00 2001
From: Javier Martinez Canillas <javier@osg.samsung.com>
Date: Mon, 6 Mar 2017 19:35:44 -0300
Subject: [PATCH] i2c: exynos5: Make transaction to retry if isn't successfully completed
After commit 7999eecb7e56 ("i2c: exynos5: fix arbitration lost handling"),
some I2C transactions are failing because the TRANSFER_DONE_AUTO field is
not set in the I2C_TRANS_STATUS register so the i2c->status value is left
to -EINVAL causing the i2c->msg_complete completion to never be signaled.
For example, when reading the time of an I2C rtc on an Exynos5800 machine:
$ cat /sys/class/rtc/rtc0/time
[ 25.924594] exynos5-hsi2c 12e10000.i2c: rx timeout
[ 65.028365] max77686-rtc max77802-rtc: Fail to read time reg(-22)
cat: /sys/class/rtc/rtc0/time: Invalid argument
To avoid this, set i2c->state to -EAGAIN if TRANSFER_DONE_AUTO is not set to
allow the I2C core to retry if the transaction wasn't successfully completed.
Signed-off-by: Javier Martinez Canillas <javier@osg.samsung.com>
---
drivers/i2c/busses/i2c-exynos5.c | 4 ++++
1 file changed, 4 insertions(+)
diff --git a/drivers/i2c/busses/i2c-exynos5.c b/drivers/i2c/busses/i2c-exynos5.c
index cbd93ce0661f..d294fb61520c 100644
--- a/drivers/i2c/busses/i2c-exynos5.c
+++ b/drivers/i2c/busses/i2c-exynos5.c
@@ -506,6 +506,10 @@ static irqreturn_t exynos5_i2c_irq(int irqno, void *dev_id)
} else if (trans_status & HSI2C_TRANS_DONE) {
i2c->trans_done = 1;
i2c->state = 0;
+ } else {
+ dev_dbg(i2c->dev, "Transaction was not completed\n");
+ i2c->state = -EAGAIN;
+ goto stop;
}
}
--
2.9.3
[2]
diff --git a/drivers/i2c/busses/i2c-exynos5.c b/drivers/i2c/busses/i2c-exynos5.c
index d294fb61520c..2b0ee9040188 100644
--- a/drivers/i2c/busses/i2c-exynos5.c
+++ b/drivers/i2c/busses/i2c-exynos5.c
@@ -774,7 +774,7 @@ static int exynos5_i2c_probe(struct platform_device *pdev)
strlcpy(i2c->adap.name, "exynos5-i2c", sizeof(i2c->adap.name));
i2c->adap.owner = THIS_MODULE;
i2c->adap.algo = &exynos5_i2c_algorithm;
- i2c->adap.retries = 3;
+ i2c->adap.retries = 5;
i2c->dev = &pdev->dev;
i2c->clk = devm_clk_get(&pdev->dev, "hsi2c");
Best regards,
--
Javier Martinez Canillas
Open Source Group
Samsung Research America
next prev parent reply other threads:[~2017-03-09 0:42 UTC|newest]
Thread overview: 12+ messages / expand[flat|nested] mbox.gz Atom feed top
[not found] <CGME20170222110438eucas1p2329ece7afd9e22ba185da64ccf4b2981@eucas1p2.samsung.com>
2017-02-22 11:04 ` [PATCH v2] i2c: exynos5: fix arbitration lost handling Andrzej Hajda
2017-02-23 6:29 ` Andi Shyti
2017-02-23 9:20 ` Andrzej Hajda
2017-02-23 12:03 ` Wolfram Sang
2017-03-08 20:05 ` Javier Martinez Canillas [this message]
2017-03-09 7:51 ` Andrzej Hajda
2017-03-09 11:03 ` Andrzej Hajda
2017-03-09 13:13 ` Javier Martinez Canillas
2017-03-09 13:49 ` Andrzej Hajda
2017-03-09 13:59 ` Javier Martinez Canillas
2017-03-09 14:23 ` Javier Martinez Canillas
2017-03-09 13:57 ` Andrzej Hajda
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=2fb11816-85e7-14f6-7b61-5d3eba6096aa@osg.samsung.com \
--to=javier@osg.samsung.com \
--cc=a.hajda@samsung.com \
--cc=andi.shyti@samsung.com \
--cc=b.zolnierkie@samsung.com \
--cc=krzk@kernel.org \
--cc=linux-i2c@vger.kernel.org \
--cc=linux-samsung-soc@vger.kernel.org \
--cc=m.szyprowski@samsung.com \
--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).