<feed xmlns='http://www.w3.org/2005/Atom'>
<title>llvm-project.git/llvm/lib/CodeGen/CodeGenPrepare.cpp, branch users/nico/python-2</title>
<subtitle>Unnamed repository; edit this file 'description' to name the repository.
</subtitle>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/'/>
<entry>
<title>[CodeGenPrepare] Filter out unrecreatable addresses from memory optimization (#143566)</title>
<updated>2025-06-28T21:30:03+00:00</updated>
<author>
<name>Evgenii Kudriashov</name>
<email>evgenii.kudriashov@intel.com</email>
</author>
<published>2025-06-28T21:30:03+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=5ffdd9480d80719dc0ff83417ae58a91c157fd79'/>
<id>5ffdd9480d80719dc0ff83417ae58a91c157fd79</id>
<content type='text'>
Follow up on #139303</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Follow up on #139303</pre>
</div>
</content>
</entry>
<entry>
<title>[DebugInfo][RemoveDIs] Remove a swathe of debug-intrinsic code (#144389)</title>
<updated>2025-06-17T14:55:14+00:00</updated>
<author>
<name>Jeremy Morse</name>
<email>jeremy.morse@sony.com</email>
</author>
<published>2025-06-17T14:55:14+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=9eb0020555fc643582b2802abb8c1bc92059c248'/>
<id>9eb0020555fc643582b2802abb8c1bc92059c248</id>
<content type='text'>
Seeing how we can't generate any debug intrinsics any more: delete a
variety of codepaths where they're handled. For the most part these are
plain deletions, in others I've tweaked comments to remain coherent, or
added a type to (what was) type-generic-lambdas.

This isn't all the DbgInfoIntrinsic call sites but it's most of the
simple scenarios.

Co-authored-by: Nikita Popov &lt;github@npopov.com&gt;</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Seeing how we can't generate any debug intrinsics any more: delete a
variety of codepaths where they're handled. For the most part these are
plain deletions, in others I've tweaked comments to remain coherent, or
added a type to (what was) type-generic-lambdas.

This isn't all the DbgInfoIntrinsic call sites but it's most of the
simple scenarios.

Co-authored-by: Nikita Popov &lt;github@npopov.com&gt;</pre>
</div>
</content>
</entry>
<entry>
<title>[DebugInfo][RemoveDIs] Delete debug-info-format flag (#143746)</title>
<updated>2025-06-12T10:51:58+00:00</updated>
<author>
<name>Jeremy Morse</name>
<email>jeremy.morse@sony.com</email>
</author>
<published>2025-06-12T10:51:58+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=97ac6483aaead89897d9bda8a12f1f4c11fad621'/>
<id>97ac6483aaead89897d9bda8a12f1f4c11fad621</id>
<content type='text'>
This flag was used to let us incrementally introduce debug records
into LLVM, however everything is now using records. It serves no
purpose now, so delete it.</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
This flag was used to let us incrementally introduce debug records
into LLVM, however everything is now using records. It serves no
purpose now, so delete it.</pre>
</div>
</content>
</entry>
<entry>
<title>[CGP] Bail out if (Base|Scaled)Reg does not dominate insert point. (#142949)</title>
<updated>2025-06-06T11:38:30+00:00</updated>
<author>
<name>Florian Hahn</name>
<email>flo@fhahn.com</email>
</author>
<published>2025-06-06T11:38:30+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=dde30a47313bf52fef02bbcb1de931a8d725659f'/>
<id>dde30a47313bf52fef02bbcb1de931a8d725659f</id>
<content type='text'>
(Base|Scaled)Reg may not dominate the chosen insert point, if there are
multiple uses of the address. Bail out if that's the case, otherwise we
will generate invalid IR.

In some cases, we could probably adjust the insert point or hoist the
(Base|Scaled)Reg.

Fixes https://github.com/llvm/llvm-project/issues/142830.

PR: https://github.com/llvm/llvm-project/pull/142949</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
(Base|Scaled)Reg may not dominate the chosen insert point, if there are
multiple uses of the address. Bail out if that's the case, otherwise we
will generate invalid IR.

In some cases, we could probably adjust the insert point or hoist the
(Base|Scaled)Reg.

Fixes https://github.com/llvm/llvm-project/issues/142830.

PR: https://github.com/llvm/llvm-project/pull/142949</pre>
</div>
</content>
</entry>
<entry>
<title>[CodeGenPrepare] Fix signed overflow (#141487)</title>
<updated>2025-06-03T07:27:25+00:00</updated>
<author>
<name>mikael-nilsson-arm</name>
<email>33650793+mikael-nilsson-arm@users.noreply.github.com</email>
</author>
<published>2025-06-03T07:27:25+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=09967917e72a3c8f02138cc0906644c0db719fbc'/>
<id>09967917e72a3c8f02138cc0906644c0db719fbc</id>
<content type='text'>
The signed addition could overflow which is undefined behavior, now the
code checks for it.</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
The signed addition could overflow which is undefined behavior, now the
code checks for it.</pre>
</div>
</content>
</entry>
<entry>
<title>Reland [llvm] add GenericFloatingPointPredicateUtils #140254 (#141065)</title>
<updated>2025-05-22T15:17:02+00:00</updated>
<author>
<name>Tim Gymnich</name>
<email>tim@gymni.ch</email>
</author>
<published>2025-05-22T15:17:02+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=571a24c314acf71db7a76449bc2bfd9a66ceb881'/>
<id>571a24c314acf71db7a76449bc2bfd9a66ceb881</id>
<content type='text'>
#140254 was previously missing 2 files in the bazel build config.</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
#140254 was previously missing 2 files in the bazel build config.</pre>
</div>
</content>
</entry>
<entry>
<title>Revert "[llvm] add GenericFloatingPointPredicateUtils (#140254)" (#140968)</title>
<updated>2025-05-21T23:31:14+00:00</updated>
<author>
<name>Kewen12</name>
<email>Kewen.Meng@amd.com</email>
</author>
<published>2025-05-21T23:31:14+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=c47a5fbb229bbdee30aeb4e5761e8c083f2fe680'/>
<id>c47a5fbb229bbdee30aeb4e5761e8c083f2fe680</id>
<content type='text'>
This reverts commit d00d74bb2564103ae3cb5ac6b6ffecf7e1cc2238. 

The PR breaks our buildbots and blocks downstream merge.</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
This reverts commit d00d74bb2564103ae3cb5ac6b6ffecf7e1cc2238. 

The PR breaks our buildbots and blocks downstream merge.</pre>
</div>
</content>
</entry>
<entry>
<title>[llvm] add GenericFloatingPointPredicateUtils (#140254)</title>
<updated>2025-05-21T21:45:31+00:00</updated>
<author>
<name>Tim Gymnich</name>
<email>tim@gymni.ch</email>
</author>
<published>2025-05-21T21:45:31+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=d00d74bb2564103ae3cb5ac6b6ffecf7e1cc2238'/>
<id>d00d74bb2564103ae3cb5ac6b6ffecf7e1cc2238</id>
<content type='text'>
add `GenericFloatingPointPredicateUtils` in order to generalize
effects of floating point comparisons on `KnownFPClass` for both IR and
MIR.

---------

Co-authored-by: Matt Arsenault &lt;arsenm2@gmail.com&gt;</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
add `GenericFloatingPointPredicateUtils` in order to generalize
effects of floating point comparisons on `KnownFPClass` for both IR and
MIR.

---------

Co-authored-by: Matt Arsenault &lt;arsenm2@gmail.com&gt;</pre>
</div>
</content>
</entry>
<entry>
<title>[CodeGenPrepare] Make sure instruction get from SunkAddrs is before MemoryInst (#139303)</title>
<updated>2025-05-15T16:27:25+00:00</updated>
<author>
<name>weiguozhi</name>
<email>57237827+weiguozhi@users.noreply.github.com</email>
</author>
<published>2025-05-15T16:27:25+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=59c6d70ed8120b8864e5f796e2bf3de5518a0ef0'/>
<id>59c6d70ed8120b8864e5f796e2bf3de5518a0ef0</id>
<content type='text'>
Function optimizeBlock may do optimizations on a block for multiple
times. In the first iteration of the loop, MemoryInst1 may generate a
sunk instruction and store it into SunkAddrs. In the second iteration of
the loop, MemoryInst2 may use the same address and then it can reuse the
sunk instruction stored in SunkAddrs, but MemoryInst2 may be before
MemoryInst1 and the corresponding sunk instruction. In order to avoid
use before def error, we need to find appropriate insert position for the
 sunk instruction.

Fixes #138208.</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Function optimizeBlock may do optimizations on a block for multiple
times. In the first iteration of the loop, MemoryInst1 may generate a
sunk instruction and store it into SunkAddrs. In the second iteration of
the loop, MemoryInst2 may use the same address and then it can reuse the
sunk instruction stored in SunkAddrs, but MemoryInst2 may be before
MemoryInst1 and the corresponding sunk instruction. In order to avoid
use before def error, we need to find appropriate insert position for the
 sunk instruction.

Fixes #138208.</pre>
</div>
</content>
</entry>
<entry>
<title>Reapply "IR: Remove uselist for constantdata (#137313)" (#138961)</title>
<updated>2025-05-08T06:00:09+00:00</updated>
<author>
<name>Matt Arsenault</name>
<email>Matthew.Arsenault@amd.com</email>
</author>
<published>2025-05-08T06:00:09+00:00</published>
<link rel='alternate' type='text/html' href='https://git.belthelziquor.com/llvm-project.git/commit/?id=9383fb23e18bb983d0024fb956a0a724ef9eb03d'/>
<id>9383fb23e18bb983d0024fb956a0a724ef9eb03d</id>
<content type='text'>
Reapply "IR: Remove uselist for constantdata (#137313)"

This reverts commit 5936c02c8b9c6d1476f7830517781ce8b6e26e75.

Fix checking uselists of constants in assume bundle queries</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Reapply "IR: Remove uselist for constantdata (#137313)"

This reverts commit 5936c02c8b9c6d1476f7830517781ce8b6e26e75.

Fix checking uselists of constants in assume bundle queries</pre>
</div>
</content>
</entry>
</feed>
