Skip to contentSkip to conversation

drivers/iio/adc/mcp3422.c:147:3-9: preceding lock on line 137 (fwd)

PATCH In review Julia Lawall 3 participants 5 messages Aug 29, 2020 – Sep 1, 2020 2 messages not in the archive IIO

lore
Share & tools
lore
#0 lore raw

PATCH 19/19 drivers/iio/adc/mcp3422.c:147:3-9: preceding lock on line 137 (fwd)

Please check whether there should be a mutex_unlock before line 147.

julia

---------- Forwarded message ----------
Date: Sun, 30 Aug 2020 04:08:59 +0800
From: kernel test robot <[email protected]>
To: kbuild(a)lists.01.org
Cc: lkp(a)intel.com, Julia Lawall <[email protected]>
Subject: [iio:fixes-togreg 19/19] drivers/iio/adc/mcp3422.c:147:3-9: preceding
lock on line 137

CC: kbuild-all(a)lists.01.org
TO: Angelo Compagnucci <[email protected]>
CC: Jonathan Cameron <[email protected]>

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git fixes-togreg
head:   ba255800f7fbb8da411c92c33b25d52970558509
commit: ba255800f7fbb8da411c92c33b25d52970558509 [19/19] iio: adc: mcp3422: fix locking scope

:::::: branch date: 3 hours ago
:::::: commit date: 3 hours ago
config: i386-randconfig-c001-20200830 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-15) 9.3.0

If you fix the issue, kindly add following tag as appropriate

coccinelle warnings: (new ones prefixed by >>)

Quoted 1 line
Quoted 1 line

drivers/iio/adc/mcp3422.c:147:3-9: preceding lock on line 137

# https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git/commit/?id=ba255800f7fbb8da411c92c33b25d52970558509
git remote add iio https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git
git fetch --no-tags iio fixes-togreg
git checkout ba255800f7fbb8da411c92c33b25d52970558509
vim +147 drivers/iio/adc/mcp3422.c

07914c84ba30e31 Angelo Compagnucci 2013-09-02  129
07914c84ba30e31 Angelo Compagnucci 2013-09-02  130  static int mcp3422_read_channel(struct mcp3422 *adc,
07914c84ba30e31 Angelo Compagnucci 2013-09-02  131  				struct iio_chan_spec const *channel, int *value)
07914c84ba30e31 Angelo Compagnucci 2013-09-02  132  {
07914c84ba30e31 Angelo Compagnucci 2013-09-02  133  	int ret;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  134  	u8 config;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  135  	u8 req_channel = channel->channel;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  136
ba255800f7fbb8d Angelo Compagnucci 2020-08-19 @137  	mutex_lock(&adc->lock);
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  138
07914c84ba30e31 Angelo Compagnucci 2013-09-02  139  	if (req_channel != MCP3422_CHANNEL(adc->config)) {
07914c84ba30e31 Angelo Compagnucci 2013-09-02  140  		config = adc->config;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  141  		config &= ~MCP3422_CHANNEL_MASK;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  142  		config |= MCP3422_CHANNEL_VALUE(req_channel);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  143  		config &= ~MCP3422_PGA_MASK;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  144  		config |= MCP3422_PGA_VALUE(adc->pga[req_channel]);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  145  		ret = mcp3422_update_config(adc, config);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  146  		if (ret < 0)
07914c84ba30e31 Angelo Compagnucci 2013-09-02 @147  			return ret;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  148  		msleep(mcp3422_read_times[MCP3422_SAMPLE_RATE(adc->config)]);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  149  	}
07914c84ba30e31 Angelo Compagnucci 2013-09-02  150
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  151  	ret = mcp3422_read(adc, value, &config);
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  152
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  153  	mutex_unlock(&adc->lock);
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  154
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  155  	return ret;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  156  }
07914c84ba30e31 Angelo Compagnucci 2013-09-02  157

:::::: The code at line 147 was first introduced by commit
:::::: 07914c84ba30e311f6bfb65b811b33a1dc094311 iio: adc: Add driver for Microchip MCP3422/3/4 high resolution ADC

:::::: TO: Angelo Compagnucci <[email protected]> :::::: CC: Jonathan Cameron <[email protected]>

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

A message that is not in the archive (sent off-list or not yet received): CA+TH9VkAo4CgCVDGvQumfePvNCg9ffwEHbqic7TsYJn4VZ3aTw@mail.gmail.com

#2 lore raw
Quoted 5 lines · Angelo Compagnucci

Hi Julia,

Il giorno sab 29 ago 2020 alle ore 22:29 Julia Lawall <[email protected]> ha scritto:

Quoted 1 line
Quoted 7 lines
Quoted 1 line

Please check whether there should be a mutex_unlock before line 147.

Having a mutex_unlock before line 147 is wrong here, cause the lock should be held for the entire reading operation. Adding an unlock before the lock means that a concurrent call can unlock the lock previously held by another call and the result ends up mixing the reading for the first call to the reading of the second call.

OK, I don't know the calling context. But you have a function where the lock is held on the failure path and is released on the success path, which seems at least a little strange. But if the calling context deals with that in a reasonable way, then ok. Maybe a comment would be useful.

julia

Quoted 87 lines

Sincerely, Angelo

Quoted 72 lines

julia

---------- Forwarded message ----------
Date: Sun, 30 Aug 2020 04:08:59 +0800
From: kernel test robot <[email protected]>
To: kbuild(a)lists.01.org
Cc: lkp(a)intel.com, Julia Lawall <[email protected]>
Subject: [iio:fixes-togreg 19/19] drivers/iio/adc/mcp3422.c:147:3-9: preceding
lock on line 137

CC: kbuild-all(a)lists.01.org
TO: Angelo Compagnucci <[email protected]>
CC: Jonathan Cameron <[email protected]>

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git fixes-togreg
head:   ba255800f7fbb8da411c92c33b25d52970558509
commit: ba255800f7fbb8da411c92c33b25d52970558509 [19/19] iio: adc: mcp3422: fix locking scope

:::::: branch date: 3 hours ago
:::::: commit date: 3 hours ago
config: i386-randconfig-c001-20200830 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-15) 9.3.0

If you fix the issue, kindly add following tag as appropriate

coccinelle warnings: (new ones prefixed by >>)

Quoted 1 line
Quoted 1 line

drivers/iio/adc/mcp3422.c:147:3-9: preceding lock on line 137

# https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git/commit/?id=ba255800f7fbb8da411c92c33b25d52970558509
git remote add iio https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git
git fetch --no-tags iio fixes-togreg
git checkout ba255800f7fbb8da411c92c33b25d52970558509
vim +147 drivers/iio/adc/mcp3422.c

07914c84ba30e31 Angelo Compagnucci 2013-09-02  129
07914c84ba30e31 Angelo Compagnucci 2013-09-02  130  static int mcp3422_read_channel(struct mcp3422 *adc,
07914c84ba30e31 Angelo Compagnucci 2013-09-02  131                              struct iio_chan_spec const *channel, int *value)
07914c84ba30e31 Angelo Compagnucci 2013-09-02  132  {
07914c84ba30e31 Angelo Compagnucci 2013-09-02  133      int ret;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  134      u8 config;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  135      u8 req_channel = channel->channel;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  136
ba255800f7fbb8d Angelo Compagnucci 2020-08-19 @137      mutex_lock(&adc->lock);
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  138
07914c84ba30e31 Angelo Compagnucci 2013-09-02  139      if (req_channel != MCP3422_CHANNEL(adc->config)) {
07914c84ba30e31 Angelo Compagnucci 2013-09-02  140              config = adc->config;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  141              config &= ~MCP3422_CHANNEL_MASK;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  142              config |= MCP3422_CHANNEL_VALUE(req_channel);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  143              config &= ~MCP3422_PGA_MASK;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  144              config |= MCP3422_PGA_VALUE(adc->pga[req_channel]);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  145              ret = mcp3422_update_config(adc, config);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  146              if (ret < 0)
07914c84ba30e31 Angelo Compagnucci 2013-09-02 @147                      return ret;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  148              msleep(mcp3422_read_times[MCP3422_SAMPLE_RATE(adc->config)]);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  149      }
07914c84ba30e31 Angelo Compagnucci 2013-09-02  150
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  151      ret = mcp3422_read(adc, value, &config);
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  152
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  153      mutex_unlock(&adc->lock);
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  154
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  155      return ret;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  156  }
07914c84ba30e31 Angelo Compagnucci 2013-09-02  157

:::::: The code at line 147 was first introduced by commit
:::::: 07914c84ba30e311f6bfb65b811b33a1dc094311 iio: adc: Add driver for Microchip MCP3422/3/4 high resolution ADC

Quoted 6 lines

:::::: TO: Angelo Compagnucci <[email protected]> :::::: CC: Jonathan Cameron <[email protected]>

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/kbuild-all(a)lists.01.org

--
Profile: http://it.linkedin.com/in/compagnucciangelo

A message that is not in the archive (sent off-list or not yet received): CA+TH9VmQq3=Kf=f72CSn2ZziKP3YP6qjsXQL1nXzS-O8FscBWw@mail.gmail.com

#4 lore raw
Quoted 24 lines · On Mon, 31 Aug 2020 11:09:53 +0200 Angelo Compagnucci…

Il giorno lun 31 ago 2020 alle ore 09:48 Julia Lawall <[email protected]> ha scritto:

Quoted 18 lines · Angelo Compagnucci
Quoted 6 lines

Hi Julia,

Il giorno sab 29 ago 2020 alle ore 22:29 Julia Lawall <[email protected]> ha scritto:

Quoted 1 line
Quoted 7 lines
Quoted 1 line

Please check whether there should be a mutex_unlock before line 147.

Having a mutex_unlock before line 147 is wrong here, cause the lock should be held for the entire reading operation. Adding an unlock before the lock means that a concurrent call can unlock the lock previously held by another call and the result ends up mixing the reading for the first call to the reading of the second call.

Quoted 3 lines

OK, I don't know the calling context. But you have a function where the lock is held on the failure path and is released on the success path, which seems at least a little strange.

Quoted 5 lines

I see.

I have to respin!

Thanks for your support!

Hi Julia, Angelo,

Please can we cc [email protected] for such reports. The fix has headed upstream. So we need to chase it with a fix asap.

Greg, would you prefer a following fix (please cc Greg directly) or to revert the patch?

3f1093d83d71 ("iio: adc: mcp3422: fix locking scope")

Sorry I missed this one. Was working on wrong computer to access the account this went to.

Jonathan

Quoted 96 lines
Quoted 93 lines

But if the calling context deals
with that in a reasonable way, then ok. Maybe a comment would be useful.

julia

Quoted 87 lines

Sincerely, Angelo

Quoted 72 lines

julia

---------- Forwarded message ----------
Date: Sun, 30 Aug 2020 04:08:59 +0800
From: kernel test robot <[email protected]>
To: [email protected]
Cc: [email protected], Julia Lawall <[email protected]>
Subject: [iio:fixes-togreg 19/19] drivers/iio/adc/mcp3422.c:147:3-9: preceding
lock on line 137

CC: [email protected]
TO: Angelo Compagnucci <[email protected]>
CC: Jonathan Cameron <[email protected]>

tree:   https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git fixes-togreg
head:   ba255800f7fbb8da411c92c33b25d52970558509
commit: ba255800f7fbb8da411c92c33b25d52970558509 [19/19] iio: adc: mcp3422: fix locking scope

:::::: branch date: 3 hours ago
:::::: commit date: 3 hours ago
config: i386-randconfig-c001-20200830 (attached as .config)
compiler: gcc-9 (Debian 9.3.0-15) 9.3.0

If you fix the issue, kindly add following tag as appropriate

coccinelle warnings: (new ones prefixed by >>)

Quoted 1 line
Quoted 1 line

drivers/iio/adc/mcp3422.c:147:3-9: preceding lock on line 137

# https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git/commit/?id=ba255800f7fbb8da411c92c33b25d52970558509
git remote add iio https://git.kernel.org/pub/scm/linux/kernel/git/jic23/iio.git
git fetch --no-tags iio fixes-togreg
git checkout ba255800f7fbb8da411c92c33b25d52970558509
vim +147 drivers/iio/adc/mcp3422.c

07914c84ba30e31 Angelo Compagnucci 2013-09-02  129
07914c84ba30e31 Angelo Compagnucci 2013-09-02  130  static int mcp3422_read_channel(struct mcp3422 *adc,
07914c84ba30e31 Angelo Compagnucci 2013-09-02  131                              struct iio_chan_spec const *channel, int *value)
07914c84ba30e31 Angelo Compagnucci 2013-09-02  132  {
07914c84ba30e31 Angelo Compagnucci 2013-09-02  133      int ret;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  134      u8 config;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  135      u8 req_channel = channel->channel;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  136
ba255800f7fbb8d Angelo Compagnucci 2020-08-19 @137      mutex_lock(&adc->lock);
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  138
07914c84ba30e31 Angelo Compagnucci 2013-09-02  139      if (req_channel != MCP3422_CHANNEL(adc->config)) {
07914c84ba30e31 Angelo Compagnucci 2013-09-02  140              config = adc->config;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  141              config &= ~MCP3422_CHANNEL_MASK;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  142              config |= MCP3422_CHANNEL_VALUE(req_channel);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  143              config &= ~MCP3422_PGA_MASK;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  144              config |= MCP3422_PGA_VALUE(adc->pga[req_channel]);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  145              ret = mcp3422_update_config(adc, config);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  146              if (ret < 0)
07914c84ba30e31 Angelo Compagnucci 2013-09-02 @147                      return ret;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  148              msleep(mcp3422_read_times[MCP3422_SAMPLE_RATE(adc->config)]);
07914c84ba30e31 Angelo Compagnucci 2013-09-02  149      }
07914c84ba30e31 Angelo Compagnucci 2013-09-02  150
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  151      ret = mcp3422_read(adc, value, &config);
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  152
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  153      mutex_unlock(&adc->lock);
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  154
ba255800f7fbb8d Angelo Compagnucci 2020-08-19  155      return ret;
07914c84ba30e31 Angelo Compagnucci 2013-09-02  156  }
07914c84ba30e31 Angelo Compagnucci 2013-09-02  157

:::::: The code at line 147 was first introduced by commit
:::::: 07914c84ba30e311f6bfb65b811b33a1dc094311 iio: adc: Add driver for Microchip MCP3422/3/4 high resolution ADC

Quoted 6 lines

:::::: TO: Angelo Compagnucci <[email protected]> :::::: CC: Jonathan Cameron <[email protected]>

---
0-DAY CI Kernel Test Service, Intel Corporation
https://lists.01.org/hyperkitty/list/[email protected]

--
Profile: http://it.linkedin.com/in/compagnucciangelo

#5 lore raw
Quoted 37 lines · Jonathan Cameron
Quoted 24 lines · On Mon, 31 Aug 2020 11:09:53 +0200 Angelo Compagnucci…

Il giorno lun 31 ago 2020 alle ore 09:48 Julia Lawall <[email protected]> ha scritto:

Quoted 18 lines · Angelo Compagnucci
Quoted 6 lines

Hi Julia,

Il giorno sab 29 ago 2020 alle ore 22:29 Julia Lawall <[email protected]> ha scritto:

Quoted 1 line
Quoted 7 lines
Quoted 1 line

Please check whether there should be a mutex_unlock before line 147.

Having a mutex_unlock before line 147 is wrong here, cause the lock should be held for the entire reading operation. Adding an unlock before the lock means that a concurrent call can unlock the lock previously held by another call and the result ends up mixing the reading for the first call to the reading of the second call.

Quoted 3 lines

OK, I don't know the calling context. But you have a function where the lock is held on the failure path and is released on the success path, which seems at least a little strange.

Quoted 5 lines

I see.

I have to respin!

Thanks for your support!

Hi Julia, Angelo,

Please can we cc [email protected] for such reports. The fix has headed upstream. So we need to chase it with a fix asap.

Quoted 7 lines

Greg, would you prefer a following fix (please cc Greg directly) or to revert the patch?

3f1093d83d71 ("iio: adc: mcp3422: fix locking scope")

Sorry I missed this one. Was working on wrong computer to access the account this went to.

A patch is always easier for me to apply than a revert is.

thanks,

greg k-h

#6 lore raw
Quoted 41 lines · On Tue, 1 Sep 2020 11:08:59 +0200 Greg KH <[email protected]> wrote:
Quoted 40 lines · Jonathan Cameron
Quoted 24 lines · On Mon, 31 Aug 2020 11:09:53 +0200 Angelo Compagnucci…

Il giorno lun 31 ago 2020 alle ore 09:48 Julia Lawall <[email protected]> ha scritto:

Quoted 18 lines · Angelo Compagnucci
Quoted 6 lines

Hi Julia,

Il giorno sab 29 ago 2020 alle ore 22:29 Julia Lawall <[email protected]> ha scritto:

Quoted 1 line
Quoted 7 lines
Quoted 1 line

Please check whether there should be a mutex_unlock before line 147.

Having a mutex_unlock before line 147 is wrong here, cause the lock should be held for the entire reading operation. Adding an unlock before the lock means that a concurrent call can unlock the lock previously held by another call and the result ends up mixing the reading for the first call to the reading of the second call.

Quoted 3 lines

OK, I don't know the calling context. But you have a function where the lock is held on the failure path and is released on the success path, which seems at least a little strange.

Quoted 5 lines

I see.

I have to respin!

Thanks for your support!

Hi Julia, Angelo,

Please can we cc [email protected] for such reports. The fix has headed upstream. So we need to chase it with a fix asap.

Greg, would you prefer a following fix (please cc Greg directly) or to revert the patch?

Quoted 6 lines
Quoted 4 lines

3f1093d83d71 ("iio: adc: mcp3422: fix locking scope")

Sorry I missed this one. Was working on wrong computer to access the account this went to.

A patch is always easier for me to apply than a revert is.

Great. Angelo replied very quickly so you should have it now (thanks Angelo!)

Thanks for your help.

Jonathan

Quoted 4 lines

thanks,

greg k-h