From: "Frank Schäfer" <fschaefer.oss@googlemail.com>
To: Mauro Carvalho Chehab <m.chehab@samsung.com>
Cc: Linux Media Mailing List <linux-media@vger.kernel.org>,
Mauro Carvalho Chehab <mchehab@infradead.org>
Subject: Re: [PATCH v4 08/22] [media] em28xx: convert i2c wait completion logic to use jiffies
Date: Mon, 06 Jan 2014 18:48:29 +0100 [thread overview]
Message-ID: <52CAEC6D.1080701@googlemail.com> (raw)
In-Reply-To: <20140105111027.06cb2aff@samsung.com>
Am 05.01.2014 14:10, schrieb Mauro Carvalho Chehab:
> Em Sun, 05 Jan 2014 12:03:51 +0100
> Frank Schäfer <fschaefer.oss@googlemail.com> escreveu:
>
>> Am 04.01.2014 11:55, schrieb Mauro Carvalho Chehab:
>>> The I2C wait completion/timeout logic currently assumes that
>>> msleep(5) will wait exaclty 5 ms. This is not true at all,
>>> as it depends on CONFIG_HZ.
>>>
>>> Convert it to use jiffies, in order to not wait for more time
>>> than needed.
>>>
>>> Signed-off-by: Mauro Carvalho Chehab <m.chehab@samsung.com>
>>> ---
>>> drivers/media/usb/em28xx/em28xx-i2c.c | 65 ++++++++++++++++++-----------------
>>> 1 file changed, 34 insertions(+), 31 deletions(-)
>>>
>>> diff --git a/drivers/media/usb/em28xx/em28xx-i2c.c b/drivers/media/usb/em28xx/em28xx-i2c.c
>>> index 9e6a11d01858..9fa7ed51e5b1 100644
>>> --- a/drivers/media/usb/em28xx/em28xx-i2c.c
>>> +++ b/drivers/media/usb/em28xx/em28xx-i2c.c
>>> @@ -26,6 +26,7 @@
>>> #include <linux/kernel.h>
>>> #include <linux/usb.h>
>>> #include <linux/i2c.h>
>>> +#include <linux/jiffies.h>
>>>
>>> #include "em28xx.h"
>>> #include "tuner-xc2028.h"
>>> @@ -48,8 +49,8 @@ MODULE_PARM_DESC(i2c_debug, "enable debug messages [i2c]");
>>> */
>>> static int em2800_i2c_send_bytes(struct em28xx *dev, u8 addr, u8 *buf, u16 len)
>>> {
>>> + unsigned long timeout = jiffies + msecs_to_jiffies(EM2800_I2C_XFER_TIMEOUT);
>>> int ret;
>>> - int write_timeout;
>>> u8 b2[6];
>>>
>>> if (len < 1 || len > 4)
>>> @@ -74,15 +75,15 @@ static int em2800_i2c_send_bytes(struct em28xx *dev, u8 addr, u8 *buf, u16 len)
>>> return (ret < 0) ? ret : -EIO;
>>> }
>>> /* wait for completion */
>>> - for (write_timeout = EM2800_I2C_XFER_TIMEOUT; write_timeout > 0;
>>> - write_timeout -= 5) {
>>> + while (time_is_after_jiffies(timeout)) {
>> AFAIU, it must be time_is_before_jiffies(timeout).
> This is tricky, but it is right.
>
> See its description at jiffies.h:
>
> /* time_is_after_jiffies(a) return true if a is after jiffies */
> #define time_is_after_jiffies(a) time_before(jiffies, a)
Urgh... yes, you are right.
I've read this, but didn't notice that we check the timeout jiffies and
not the current jiffies. :/
Sorry for the noise.
>
>>> ret = dev->em28xx_read_reg(dev, 0x05);
>>> - if (ret == 0x80 + len - 1) {
>>> + if (ret == 0x80 + len - 1)
>>> return len;
>>> - } else if (ret == 0x94 + len - 1) {
>>> + if (ret == 0x94 + len - 1) {
>>> em28xx_warn("R05 returned 0x%02x: I2C timeout", ret);
>>> return -ENODEV;
>>> - } else if (ret < 0) {
>>> + }
>>> + if (ret < 0) {
>>> em28xx_warn("failed to get i2c transfer status from bridge register (error=%i)\n",
>>> ret);
>>> return ret;
>>> @@ -99,9 +100,9 @@ static int em2800_i2c_send_bytes(struct em28xx *dev, u8 addr, u8 *buf, u16 len)
>>> */
>>> static int em2800_i2c_recv_bytes(struct em28xx *dev, u8 addr, u8 *buf, u16 len)
>>> {
>>> + unsigned long timeout = jiffies + msecs_to_jiffies(EM2800_I2C_XFER_TIMEOUT);
>>> u8 buf2[4];
>>> int ret;
>>> - int read_timeout;
>>> int i;
>>>
>>> if (len < 1 || len > 4)
>>> @@ -118,15 +119,15 @@ static int em2800_i2c_recv_bytes(struct em28xx *dev, u8 addr, u8 *buf, u16 len)
>>> }
>>>
>>> /* wait for completion */
>>> - for (read_timeout = EM2800_I2C_XFER_TIMEOUT; read_timeout > 0;
>>> - read_timeout -= 5) {
>>> + while (time_is_after_jiffies(timeout)) {
>> The same here...
>>
>>> ret = dev->em28xx_read_reg(dev, 0x05);
>>> - if (ret == 0x84 + len - 1) {
>>> + if (ret == 0x84 + len - 1)
>>> break;
>>> - } else if (ret == 0x94 + len - 1) {
>>> + if (ret == 0x94 + len - 1) {
>>> em28xx_warn("R05 returned 0x%02x: I2C timeout", ret);
>>> return -ENODEV;
>>> - } else if (ret < 0) {
>>> + }
>>> + if (ret < 0) {
>>> em28xx_warn("failed to get i2c transfer status from bridge register (error=%i)\n",
>>> ret);
>>> return ret;
>>> @@ -170,7 +171,8 @@ static int em2800_i2c_check_for_device(struct em28xx *dev, u8 addr)
>>> static int em28xx_i2c_send_bytes(struct em28xx *dev, u16 addr, u8 *buf,
>>> u16 len, int stop)
>>> {
>>> - int write_timeout, ret;
>>> + unsigned long timeout = jiffies + msecs_to_jiffies(EM2800_I2C_XFER_TIMEOUT);
>>> + int ret;
>>>
>>> if (len < 1 || len > 64)
>>> return -EOPNOTSUPP;
>>> @@ -193,17 +195,18 @@ static int em28xx_i2c_send_bytes(struct em28xx *dev, u16 addr, u8 *buf,
>>> }
>>> }
>>>
>>> - /* Check success of the i2c operation */
>>> - for (write_timeout = EM2800_I2C_XFER_TIMEOUT; write_timeout > 0;
>>> - write_timeout -= 5) {
>>> + /* wait for completion */
>>> + while (time_is_after_jiffies(timeout)) {
>> ... and here.
>>
>>> ret = dev->em28xx_read_reg(dev, 0x05);
>>> - if (ret == 0) { /* success */
>>> + if (ret == 0) /* success */
>>> return len;
>>> - } else if (ret == 0x10) {
>>> - em28xx_warn("I2C transfer timeout on writing to addr 0x%02x", addr);
>>> + if (ret == 0x10) {
>>> + em28xx_warn("I2C transfer timeout on writing to addr 0x%02x",
>>> + addr);
>>> return -ENODEV;
>>> - } else if (ret < 0) {
>>> - em28xx_warn("failed to read i2c transfer status from bridge (error=%i)\n",
>>> + }
>>> + if (ret < 0) {
>>> + em28xx_warn("failed to get i2c transfer status from bridge register (error=%i)\n",
>>> ret);
>>> return ret;
>>> }
>>> @@ -214,6 +217,7 @@ static int em28xx_i2c_send_bytes(struct em28xx *dev, u16 addr, u8 *buf,
>>> * (even with high payload) ...
>>> */
>>> }
>>> +
>>> em28xx_warn("write to i2c device at 0x%x timed out\n", addr);
>>> return -EIO;
>>> }
>>> @@ -251,21 +255,20 @@ static int em28xx_i2c_recv_bytes(struct em28xx *dev, u16 addr, u8 *buf, u16 len)
>>>
>>> /* Check success of the i2c operation */
>>> ret = dev->em28xx_read_reg(dev, 0x05);
>>> + if (ret == 0) /* success */
>>> + return len;
>>> if (ret < 0) {
>>> - em28xx_warn("failed to read i2c transfer status from bridge (error=%i)\n",
>>> + em28xx_warn("failed to get i2c transfer status from bridge register (error=%i)\n",
>>> ret);
>>> return ret;
>>> }
>>> - if (ret > 0) {
>>> - if (ret == 0x10) {
>>> - em28xx_warn("I2C transfer timeout on read from addr 0x%02x", addr);
>>> - return -ENODEV;
>>> - } else {
>>> - em28xx_warn("unknown i2c error (status=%i)\n", ret);
>>> - return -EIO;
>>> - }
>>> + if (ret == 0x10) {
>>> + em28xx_warn("I2C transfer timeout on read from addr 0x%02x", addr);
>>> + return -ENODEV;
>>> }
>>> - return len;
>>> +
>>> + em28xx_warn("unknown i2c error (status=%i)\n", ret);
>>> + return -EIO;
>>> }
>>>
>>> /*
>
next prev parent reply other threads:[~2014-01-06 17:47 UTC|newest]
Thread overview: 80+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-01-04 10:55 [PATCH v4 00/22] em28xx: split analog part into a separate module Mauro Carvalho Chehab
2014-01-04 10:55 ` [PATCH v4 01/22] [media] em28xx: move some video-specific functions to em28xx-video Mauro Carvalho Chehab
2014-01-05 10:11 ` Frank Schäfer
2014-01-05 13:28 ` Mauro Carvalho Chehab
2014-01-04 10:55 ` [PATCH v4 02/22] [media] em28xx: some cosmetic changes Mauro Carvalho Chehab
2014-01-05 10:13 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 03/22] [media] em28xx: move analog-specific init to em28xx-video Mauro Carvalho Chehab
2014-01-05 10:26 ` Frank Schäfer
2014-01-05 14:40 ` Mauro Carvalho Chehab
2014-01-06 21:28 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 04/22] [media] em28xx: make em28xx-video to be a separate module Mauro Carvalho Chehab
2014-01-05 10:47 ` Frank Schäfer
2014-01-05 12:56 ` Mauro Carvalho Chehab
2014-01-05 15:18 ` Mauro Carvalho Chehab
2014-01-06 21:35 ` Frank Schäfer
2014-01-06 17:38 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 05/22] [media] em28xx: initialize analog I2C devices at the right place Mauro Carvalho Chehab
2014-01-05 10:48 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 06/22] [media] em28xx: add warn messages for timeout Mauro Carvalho Chehab
2014-01-05 10:51 ` Frank Schäfer
2014-01-05 13:05 ` Mauro Carvalho Chehab
2014-01-05 13:25 ` Mauro Carvalho Chehab
2014-01-04 10:55 ` [PATCH v4 07/22] [media] em28xx: improve extension information messages Mauro Carvalho Chehab
2014-01-05 10:55 ` Frank Schäfer
2014-01-05 13:08 ` Mauro Carvalho Chehab
2014-01-05 15:31 ` Mauro Carvalho Chehab
2014-01-06 17:44 ` Frank Schäfer
2014-01-06 18:17 ` Mauro Carvalho Chehab
2014-01-04 10:55 ` [PATCH v4 08/22] [media] em28xx: convert i2c wait completion logic to use jiffies Mauro Carvalho Chehab
2014-01-05 11:03 ` Frank Schäfer
2014-01-05 13:10 ` Mauro Carvalho Chehab
2014-01-06 17:48 ` Frank Schäfer [this message]
2014-01-04 10:55 ` [PATCH v4 09/22] [media] tvp5150: make read operations atomic Mauro Carvalho Chehab
2014-01-05 11:07 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 10/22] [media] tuner-xc2028: remove unused code Mauro Carvalho Chehab
2014-01-05 11:07 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 11/22] [media] em28xx: check if a device has audio earlier Mauro Carvalho Chehab
2014-01-05 11:12 ` Frank Schäfer
2014-01-05 13:22 ` Mauro Carvalho Chehab
2014-01-04 10:55 ` [PATCH v4 12/22] [media] em28xx: properly implement AC97 wait code Mauro Carvalho Chehab
2014-01-05 11:19 ` Frank Schäfer
2014-01-05 13:20 ` Mauro Carvalho Chehab
2014-01-05 15:44 ` Mauro Carvalho Chehab
2014-01-07 16:50 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 13/22] [media] em28xx: initialize audio latter Mauro Carvalho Chehab
2014-01-05 11:29 ` Frank Schäfer
2014-01-05 13:17 ` Mauro Carvalho Chehab
2014-01-07 17:00 ` Frank Schäfer
2014-01-08 14:29 ` Mauro Carvalho Chehab
2014-01-04 10:55 ` [PATCH v4 14/22] [media] em28xx: unify module version Mauro Carvalho Chehab
2014-01-05 11:33 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 15/22] [media] em28xx: Fix em28xx deplock Mauro Carvalho Chehab
2014-01-05 11:38 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 16/22] [media] em28xx: use a better value for I2C timeouts Mauro Carvalho Chehab
2014-01-05 20:38 ` Frank Schäfer
2014-01-05 20:57 ` Mauro Carvalho Chehab
2014-01-07 17:15 ` Frank Schäfer
2014-01-08 14:39 ` Mauro Carvalho Chehab
2014-01-04 10:55 ` [PATCH v4 17/22] [media] em28xx-i2c: Fix error code for I2C error transfers Mauro Carvalho Chehab
2014-01-05 20:40 ` Frank Schäfer
2014-01-06 9:55 ` Mauro Carvalho Chehab
2014-01-07 17:28 ` Frank Schäfer
2014-01-08 11:55 ` Mauro Carvalho Chehab
2014-01-08 19:37 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 18/22] [media] em28xx: don't return -ENODEV for I2C xfer errors Mauro Carvalho Chehab
2014-01-05 20:49 ` Frank Schäfer
2014-01-06 10:37 ` Mauro Carvalho Chehab
2014-01-04 10:55 ` [PATCH v4 19/22] [media] em28xx: cleanup I2C debug messages Mauro Carvalho Chehab
2014-01-05 20:54 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 20/22] [media] em28xx: use usb_alloc_coherent() for audio Mauro Carvalho Chehab
2014-01-05 20:57 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 21/22] [media] em28xx-audio: allocate URBs at device driver init Mauro Carvalho Chehab
2014-01-05 21:02 ` Frank Schäfer
2014-01-05 21:25 ` Mauro Carvalho Chehab
2014-01-06 16:25 ` Mauro Carvalho Chehab
2014-01-07 17:03 ` Frank Schäfer
2014-01-08 14:10 ` Mauro Carvalho Chehab
2014-01-08 19:14 ` Frank Schäfer
2014-01-04 10:55 ` [PATCH v4 22/22] [media] em28xx: retry read operation if it fails Mauro Carvalho Chehab
2014-01-05 21:06 ` Frank Schäfer
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=52CAEC6D.1080701@googlemail.com \
--to=fschaefer.oss@googlemail.com \
--cc=linux-media@vger.kernel.org \
--cc=m.chehab@samsung.com \
--cc=mchehab@infradead.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.