HTML code for sharing source code

HTML Copy To Clipboarddone!

<div class="highlight" style="background: #f8f8f8"><pre style="line-height: 125%"><div id="lc1">We&#39;re seeing a intermittient though quite frequent bug in our kernel:
</div><div id="lc2">all of the sudden N516 stops responding to keys and quickly drains
</div><div id="lc3">battery (due to being unable to enter sleep mode). We revealed the
</div><div id="lc4">following pattern in logs repeating infinitely:
</div><div id="lc5">
</div><div id="lc6">&lt;7&gt;[13331.730000] i2c-adapter i2c-0: jz47xx_xfer: processing 1 messages:
</div><div id="lc7">&lt;7&gt;[13331.730000] i2c-adapter i2c-0:  #0: reading 1 byte from 0x54
</div><div id="lc8">&lt;7&gt;[13331.730000] i2c-adapter i2c-0: xfer_read
</div><div id="lc9">&lt;7&gt;[13331.830000] i2c-adapter i2c-0: xfer_read: 114: timeout
</div><div id="lc10">&lt;7&gt;[13331.830000] i2c-adapter i2c-0: xfer_read: Received ACK
</div><div id="lc11">&lt;7&gt;[13331.830000] i2c-adapter i2c-0: transfer complete
</div><div id="lc12">&lt;7&gt;[13331.830000] n516-keys 0-0054: msg: 0xa9
</div><div id="lc13">
</div><div id="lc14">Did we do anything wrong so controller went mad?
</div><div id="lc15">
</div><div id="lc16">Is it possible completely reset the i2c controller as a fail-safe
</div><div id="lc17">measure in such situations? As you could see from the code (especially
</div><div id="lc18">i2c-jz47xx.c::xfer_read) we are disabling i2c after every transaction.
</div><div id="lc19">
</div><div id="lc20">For the reference, there are links to a source code for jz47x0
</div><div id="lc21">controller driver and n516 binding:
</div><div id="lc22">
</div><div id="lc23">http://git.openinkpot.org/linux-2.6.git/?a=blob;f=drivers/i2c/busses/i2c-jz47xx.c
</div><div id="lc24">http://git.openinkpot.org/linux-2.6.git/?a=blob;f=arch/mips/jz4740/i2c.c
</div><div id="lc25">http://git.openinkpot.org/linux-2.6.git/?a=blob;f=drivers/i2c/chips/n516-lpc.c
</div></pre></div>
<p style="padding: 0; margin: 2em 0 1em 0; font-size: 0.8em; color: #666;"><span style="padding: 0 1em; background-color: #f5f5f5;"> Colored with <a style="color: #333" href="http://dumpz.org">dumpz.org</a></span></p>

BBCODE Copy To Clipboarddone!

We're seeing a intermittient though quite frequent bug in our kernel:
all of the sudden N516 stops responding to keys and quickly drains
battery (due to being unable to enter sleep mode). We revealed the
following pattern in logs repeating infinitely:

<7>[13331.730000] i2c-adapter i2c-0: jz47xx_xfer: processing 1 messages:
<7>[13331.730000] i2c-adapter i2c-0:  #0: reading 1 byte from 0x54
<7>[13331.730000] i2c-adapter i2c-0: xfer_read
<7>[13331.830000] i2c-adapter i2c-0: xfer_read: 114: timeout
<7>[13331.830000] i2c-adapter i2c-0: xfer_read: Received ACK
<7>[13331.830000] i2c-adapter i2c-0: transfer complete
<7>[13331.830000] n516-keys 0-0054: msg: 0xa9

Did we do anything wrong so controller went mad?

Is it possible completely reset the i2c controller as a fail-safe
measure in such situations? As you could see from the code (especially
i2c-jz47xx.c::xfer_read) we are disabling i2c after every transaction.

For the reference, there are links to a source code for jz47x0
controller driver and n516 binding:

--------------------------------------
Colored with http://dumpz.org

Light HTML Copy To Clipboarddone!

That method generates HTML with CSS classes. Hence you'll need the style sheet file

<div class="highlight"><pre><div id="lc1">We&#39;re seeing a intermittient though quite frequent bug in our kernel:
</div><div id="lc2">all of the sudden N516 stops responding to keys and quickly drains
</div><div id="lc3">battery (due to being unable to enter sleep mode). We revealed the
</div><div id="lc4">following pattern in logs repeating infinitely:
</div><div id="lc5">
</div><div id="lc6">&lt;7&gt;[13331.730000] i2c-adapter i2c-0: jz47xx_xfer: processing 1 messages:
</div><div id="lc7">&lt;7&gt;[13331.730000] i2c-adapter i2c-0:  #0: reading 1 byte from 0x54
</div><div id="lc8">&lt;7&gt;[13331.730000] i2c-adapter i2c-0: xfer_read
</div><div id="lc9">&lt;7&gt;[13331.830000] i2c-adapter i2c-0: xfer_read: 114: timeout
</div><div id="lc10">&lt;7&gt;[13331.830000] i2c-adapter i2c-0: xfer_read: Received ACK
</div><div id="lc11">&lt;7&gt;[13331.830000] i2c-adapter i2c-0: transfer complete
</div><div id="lc12">&lt;7&gt;[13331.830000] n516-keys 0-0054: msg: 0xa9
</div><div id="lc13">
</div><div id="lc14">Did we do anything wrong so controller went mad?
</div><div id="lc15">
</div><div id="lc16">Is it possible completely reset the i2c controller as a fail-safe
</div><div id="lc17">measure in such situations? As you could see from the code (especially
</div><div id="lc18">i2c-jz47xx.c::xfer_read) we are disabling i2c after every transaction.
</div><div id="lc19">
</div><div id="lc20">For the reference, there are links to a source code for jz47x0
</div><div id="lc21">controller driver and n516 binding:
</div><div id="lc22">
</div><div id="lc23">http://git.openinkpot.org/linux-2.6.git/?a=blob;f=drivers/i2c/busses/i2c-jz47xx.c
</div><div id="lc24">http://git.openinkpot.org/linux-2.6.git/?a=blob;f=arch/mips/jz4740/i2c.c
</div><div id="lc25">http://git.openinkpot.org/linux-2.6.git/?a=blob;f=drivers/i2c/chips/n516-lpc.c
</div></pre></div>
<p style="padding: 0; margin: 2em 0 1em 0; font-size: 0.8em; color: #666;"><span style="padding: 0 1em; background-color: #f5f5f5;"> Colored with <a style="color: #333" href="http://dumpz.org">dumpz.org</a></span></p>

Habrahabr.ru HTML Copy To Clipboarddone!

Use that HTML for publications on habrahabr.ru

<blockquote><code>We&#39;re&nbsp;seeing&nbsp;a&nbsp;intermittient&nbsp;though&nbsp;quite&nbsp;frequent&nbsp;bug&nbsp;in&nbsp;our&nbsp;kernel:
all&nbsp;of&nbsp;the&nbsp;sudden&nbsp;N516&nbsp;stops&nbsp;responding&nbsp;to&nbsp;keys&nbsp;and&nbsp;quickly&nbsp;drains
battery&nbsp;(due&nbsp;to&nbsp;being&nbsp;unable&nbsp;to&nbsp;enter&nbsp;sleep&nbsp;mode).&nbsp;We&nbsp;revealed&nbsp;the
following&nbsp;pattern&nbsp;in&nbsp;logs&nbsp;repeating&nbsp;infinitely:

&lt;7&gt;[13331.730000]&nbsp;i2c-adapter&nbsp;i2c-0:&nbsp;jz47xx_xfer:&nbsp;processing&nbsp;1&nbsp;messages:
&lt;7&gt;[13331.730000]&nbsp;i2c-adapter&nbsp;i2c-0:&nbsp;&nbsp;#0:&nbsp;reading&nbsp;1&nbsp;byte&nbsp;from&nbsp;0x54
&lt;7&gt;[13331.730000]&nbsp;i2c-adapter&nbsp;i2c-0:&nbsp;xfer_read
&lt;7&gt;[13331.830000]&nbsp;i2c-adapter&nbsp;i2c-0:&nbsp;xfer_read:&nbsp;114:&nbsp;timeout
&lt;7&gt;[13331.830000]&nbsp;i2c-adapter&nbsp;i2c-0:&nbsp;xfer_read:&nbsp;Received&nbsp;ACK
&lt;7&gt;[13331.830000]&nbsp;i2c-adapter&nbsp;i2c-0:&nbsp;transfer&nbsp;complete
&lt;7&gt;[13331.830000]&nbsp;n516-keys&nbsp;0-0054:&nbsp;msg:&nbsp;0xa9

Did&nbsp;we&nbsp;do&nbsp;anything&nbsp;wrong&nbsp;so&nbsp;controller&nbsp;went&nbsp;mad?

Is&nbsp;it&nbsp;possible&nbsp;completely&nbsp;reset&nbsp;the&nbsp;i2c&nbsp;controller&nbsp;as&nbsp;a&nbsp;fail-safe
measure&nbsp;in&nbsp;such&nbsp;situations?&nbsp;As&nbsp;you&nbsp;could&nbsp;see&nbsp;from&nbsp;the&nbsp;code&nbsp;(especially
i2c-jz47xx.c::xfer_read)&nbsp;we&nbsp;are&nbsp;disabling&nbsp;i2c&nbsp;after&nbsp;every&nbsp;transaction.

For&nbsp;the&nbsp;reference,&nbsp;there&nbsp;are&nbsp;links&nbsp;to&nbsp;a&nbsp;source&nbsp;code&nbsp;for&nbsp;jz47x0
controller&nbsp;driver&nbsp;and&nbsp;n516&nbsp;binding:

http://git.openinkpot.org/linux-2.6.git/?a=blob;f=drivers/i2c/busses/i2c-jz47xx.c
http://git.openinkpot.org/linux-2.6.git/?a=blob;f=arch/mips/jz4740/i2c.c
http://git.openinkpot.org/linux-2.6.git/?a=blob;f=drivers/i2c/chips/n516-lpc.c
</code></blockquote><p><font color="gray">Colored with <a href="http://dumpz.org">dumpz.org</a></p>