<feed xmlns='http://www.w3.org/2005/Atom'>
<title>linux-toradex.git/arch/powerpc/kernel/align.c, branch v3.12.45</title>
<subtitle>Linux kernel for Apalis and Colibri modules</subtitle>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/'/>
<entry>
<title>powerpc: Never handle VSX alignment exceptions from kernel</title>
<updated>2013-08-27T04:44:26+00:00</updated>
<author>
<name>Anton Blanchard</name>
<email>anton@samba.org</email>
</author>
<published>2013-08-20T10:30:07+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=5c2e08231b68a3c8082716a7ed4e972dde406e4a'/>
<id>5c2e08231b68a3c8082716a7ed4e972dde406e4a</id>
<content type='text'>
The VSX alignment handler needs to write out the existing VSX
state to memory before operating on it (flush_vsx_to_thread()).
If we take a VSX alignment exception in the kernel bad things
will happen. It looks like we could write the kernel state out
to the user process, or we could handle the kernel exception
using data from the user process (depending if MSR_VSX is set
or not).

Worse still, if the code to read or write the VSX state causes an
alignment exception, we will recurse forever. I ended up with
hundreds of megabytes of kernel stack to look through as a result.

Floating point and SPE code have similar issues but already include
a user check. Add the same check to emulate_vsx().

With this patch any unaligned VSX loads and stores in the kernel
will show up as a clear oops rather than silent corruption of
kernel or userspace VSX state, or worse, corruption of a potentially
unlimited amount of kernel memory.

Signed-off-by: Anton Blanchard &lt;anton@samba.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
The VSX alignment handler needs to write out the existing VSX
state to memory before operating on it (flush_vsx_to_thread()).
If we take a VSX alignment exception in the kernel bad things
will happen. It looks like we could write the kernel state out
to the user process, or we could handle the kernel exception
using data from the user process (depending if MSR_VSX is set
or not).

Worse still, if the code to read or write the VSX state causes an
alignment exception, we will recurse forever. I ended up with
hundreds of megabytes of kernel stack to look through as a result.

Floating point and SPE code have similar issues but already include
a user check. Add the same check to emulate_vsx().

With this patch any unaligned VSX loads and stores in the kernel
will show up as a clear oops rather than silent corruption of
kernel or userspace VSX state, or worse, corruption of a potentially
unlimited amount of kernel memory.

Signed-off-by: Anton Blanchard &lt;anton@samba.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>powerpc: Handle unaligned ldbrx/stdbrx</title>
<updated>2013-08-14T01:50:20+00:00</updated>
<author>
<name>Anton Blanchard</name>
<email>anton@samba.org</email>
</author>
<published>2013-08-06T16:01:19+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=230aef7a6a23b6166bd4003bfff5af23c9bd381f'/>
<id>230aef7a6a23b6166bd4003bfff5af23c9bd381f</id>
<content type='text'>
Normally when we haven't implemented an alignment handler for
a load or store instruction the process will be terminated.

The alignment handler uses the DSISR (or a pseudo one) to locate
the right handler. Unfortunately ldbrx and stdbrx overlap lfs and
stfs so we incorrectly think ldbrx is an lfs and stdbrx is an
stfs.

This bug is particularly nasty - instead of terminating the
process we apply an incorrect fixup and continue on.

With more and more overlapping instructions we should stop
creating a pseudo DSISR and index using the instruction directly,
but for now add a special case to catch ldbrx/stdbrx.

Signed-off-by: Anton Blanchard &lt;anton@samba.org&gt;
Cc: &lt;stable@vger.kernel.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Normally when we haven't implemented an alignment handler for
a load or store instruction the process will be terminated.

The alignment handler uses the DSISR (or a pseudo one) to locate
the right handler. Unfortunately ldbrx and stdbrx overlap lfs and
stfs so we incorrectly think ldbrx is an lfs and stdbrx is an
stfs.

This bug is particularly nasty - instead of terminating the
process we apply an incorrect fixup and continue on.

With more and more overlapping instructions we should stop
creating a pseudo DSISR and index using the instruction directly,
but for now add a special case to catch ldbrx/stdbrx.

Signed-off-by: Anton Blanchard &lt;anton@samba.org&gt;
Cc: &lt;stable@vger.kernel.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>Disintegrate asm/system.h for PowerPC</title>
<updated>2012-03-28T17:30:02+00:00</updated>
<author>
<name>David Howells</name>
<email>dhowells@redhat.com</email>
</author>
<published>2012-03-28T17:30:02+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=ae3a197e3d0bfe3f4bf1693723e82dc018c096f3'/>
<id>ae3a197e3d0bfe3f4bf1693723e82dc018c096f3</id>
<content type='text'>
Disintegrate asm/system.h for PowerPC.

Signed-off-by: David Howells &lt;dhowells@redhat.com&gt;
Acked-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
cc: linuxppc-dev@lists.ozlabs.org
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Disintegrate asm/system.h for PowerPC.

Signed-off-by: David Howells &lt;dhowells@redhat.com&gt;
Acked-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
cc: linuxppc-dev@lists.ozlabs.org
</pre>
</div>
</content>
</entry>
<entry>
<title>powerpc: Remove fpscr use from [kvm_]cvt_{fd,df}</title>
<updated>2010-09-02T04:07:32+00:00</updated>
<author>
<name>Andreas Schwab</name>
<email>schwab@linux-m68k.org</email>
</author>
<published>2010-08-21T11:43:20+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=05d77ac90c0d260ae18decd70507dc4f5b71a2cb'/>
<id>05d77ac90c0d260ae18decd70507dc4f5b71a2cb</id>
<content type='text'>
Neither lfs nor stfs touch the fpscr, so remove the restore/save of it
around them.

Signed-off-by: Andreas Schwab &lt;schwab@linux-m68k.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Neither lfs nor stfs touch the fpscr, so remove the restore/save of it
around them.

Signed-off-by: Andreas Schwab &lt;schwab@linux-m68k.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>powerpc: Handle VSX alignment faults correctly in little-endian mode</title>
<updated>2009-12-18T03:55:43+00:00</updated>
<author>
<name>Neil Campbell</name>
<email>neilc@linux.vnet.ibm.com</email>
</author>
<published>2009-12-14T04:08:57+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=bb7f20b1c639606def3b91f4e4aca6daeee5d80a'/>
<id>bb7f20b1c639606def3b91f4e4aca6daeee5d80a</id>
<content type='text'>
This patch fixes the handling of VSX alignment faults in little-endian
mode (the current code assumes the processor is in big-endian mode).

The patch also makes the handlers clear the top 8 bytes of the register
when handling an 8 byte VSX load.

This is based on 2.6.32.

Signed-off-by: Neil Campbell &lt;neilc@linux.vnet.ibm.com&gt;
Cc: &lt;stable@kernel.org&gt;
Acked-by: Michael Neuling &lt;mikey@neuling.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
This patch fixes the handling of VSX alignment faults in little-endian
mode (the current code assumes the processor is in big-endian mode).

The patch also makes the handlers clear the top 8 bytes of the register
when handling an 8 byte VSX load.

This is based on 2.6.32.

Signed-off-by: Neil Campbell &lt;neilc@linux.vnet.ibm.com&gt;
Cc: &lt;stable@kernel.org&gt;
Acked-by: Michael Neuling &lt;mikey@neuling.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>powerpc: Create PPC_WARN_ALIGNMENT to match PPC_WARN_EMULATED</title>
<updated>2009-10-28T05:13:03+00:00</updated>
<author>
<name>Anton Blanchard</name>
<email>anton@samba.org</email>
</author>
<published>2009-10-27T18:46:55+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=eecff81d1fcda22cd0029d11fe2a71dceed11dad'/>
<id>eecff81d1fcda22cd0029d11fe2a71dceed11dad</id>
<content type='text'>
perf_event wants a separate event for alignment and emulation faults,
so create another emulation event.  This will make it easy to hook in
perf_event at one spot.

We pass in regs which will be required for these events.

Signed-off-by: Anton Blanchard &lt;anton@samba.org&gt;
Signed-off-by: Paul Mackerras &lt;paulus@samba.org&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
perf_event wants a separate event for alignment and emulation faults,
so create another emulation event.  This will make it easy to hook in
perf_event at one spot.

We pass in regs which will be required for these events.

Signed-off-by: Anton Blanchard &lt;anton@samba.org&gt;
Signed-off-by: Paul Mackerras &lt;paulus@samba.org&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>powerpc: Keep track of emulated instructions</title>
<updated>2009-05-21T05:44:26+00:00</updated>
<author>
<name>Geert Uytterhoeven</name>
<email>Geert.Uytterhoeven@sonycom.com</email>
</author>
<published>2009-05-18T02:10:05+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=80947e7c99c29ce3a78bdc1933b310468455a82f'/>
<id>80947e7c99c29ce3a78bdc1933b310468455a82f</id>
<content type='text'>
If CONFIG_PPC_EMULATED_STATS is enabled, make available counters for the
various classes of emulated instructions under
/sys/kernel/debug/powerpc/emulated_instructions/ (assumed debugfs is mounted on
/sys/kernel/debug).  Optionally (controlled by
/sys/kernel/debug/powerpc/emulated_instructions/do_warn), rate-limited warnings
can be printed to the console when instructions are emulated.

Signed-off-by: Geert Uytterhoeven &lt;Geert.Uytterhoeven@sonycom.com&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
If CONFIG_PPC_EMULATED_STATS is enabled, make available counters for the
various classes of emulated instructions under
/sys/kernel/debug/powerpc/emulated_instructions/ (assumed debugfs is mounted on
/sys/kernel/debug).  Optionally (controlled by
/sys/kernel/debug/powerpc/emulated_instructions/do_warn), rate-limited warnings
can be printed to the console when instructions are emulated.

Signed-off-by: Geert Uytterhoeven &lt;Geert.Uytterhoeven@sonycom.com&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>powerpc: Fix load/store float double alignment handler</title>
<updated>2009-02-23T04:53:05+00:00</updated>
<author>
<name>Michael Neuling</name>
<email>mikey@neuling.org</email>
</author>
<published>2009-02-19T18:52:20+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=553631e25f238de3a8085d2daf9cd4dcd96f8573'/>
<id>553631e25f238de3a8085d2daf9cd4dcd96f8573</id>
<content type='text'>
When we introduced VSX, we changed the way FPRs are stored in the
thread_struct.  Unfortunately we missed the load/store float double
alignment handler code when updating how we access FPRs in the
thread_struct.

Below fixes this and merges the little/big endian case.

Signed-off-by: Michael Neuling &lt;mikey@neuling.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
When we introduced VSX, we changed the way FPRs are stored in the
thread_struct.  Unfortunately we missed the load/store float double
alignment handler code when updating how we access FPRs in the
thread_struct.

Below fixes this and merges the little/big endian case.

Signed-off-by: Michael Neuling &lt;mikey@neuling.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>powerpc: Add alignment handler for new lfiwzx instruction</title>
<updated>2009-02-23T04:53:04+00:00</updated>
<author>
<name>Michael Neuling</name>
<email>mikey@neuling.org</email>
</author>
<published>2009-02-19T18:51:37+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=545bba18247067bb63c94e042bed90599d08151b'/>
<id>545bba18247067bb63c94e042bed90599d08151b</id>
<content type='text'>
lfiwzx is a new floating point load instruction in 2.06 that needs an
alignment handler for Linux.

Turns out to be the worlds easiest handler to add.

Signed-off-by: Michael Neuling &lt;mikey@neuling.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
lfiwzx is a new floating point load instruction in 2.06 that needs an
alignment handler for Linux.

Turns out to be the worlds easiest handler to add.

Signed-off-by: Michael Neuling &lt;mikey@neuling.org&gt;
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>powerpc/vsx: Fix VSX alignment handler for regs 32-63</title>
<updated>2009-02-13T05:37:45+00:00</updated>
<author>
<name>Michael Neuling</name>
<email>mikey@neuling.org</email>
</author>
<published>2009-02-12T19:08:58+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=26456dcfb8d8e43b1b64b2a14710694cf7a72f05'/>
<id>26456dcfb8d8e43b1b64b2a14710694cf7a72f05</id>
<content type='text'>
Fix the VSX alignment handler for VSX registers &gt; 32.  32-63 are stored
in the VMX part of the thread_struct not the FPR part.

Signed-off-by: Michael Neuling &lt;mikey@neuling.org&gt;
CC: stable@kernel.org (2.6.27 &amp; .28 please)
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Fix the VSX alignment handler for VSX registers &gt; 32.  32-63 are stored
in the VMX part of the thread_struct not the FPR part.

Signed-off-by: Michael Neuling &lt;mikey@neuling.org&gt;
CC: stable@kernel.org (2.6.27 &amp; .28 please)
Signed-off-by: Benjamin Herrenschmidt &lt;benh@kernel.crashing.org&gt;
</pre>
</div>
</content>
</entry>
</feed>
