public inbox for git-commits@fedoraproject.org
help / color / mirror / Atom feed
From: Androniychuk, Pavel <pavel.androniychuk@intel.com>
To: git-commits@fedoraproject.org
Subject: [rpms/intel-igc] epel10.2: Add patch for LLVM15 segfault
Date: Fri, 12 Jun 2026 23:37:52 GMT [thread overview]
Message-ID: <178130747248.1.11157231997824707851.rpms-intel-igc-96faceb1cb09@fedoraproject.org> (raw)
A new commit has been pushed.
Repo : rpms/intel-igc
Branch : epel10.2
Commit : 96faceb1cb090c9902af4593cb8482b5fec8c109
Author : Androniychuk, Pavel <pavel.androniychuk@intel.com>
Date : 2026-06-09T10:12:29-07:00
Stats : +32/-0 in 2 file(s)
URL : https://src.fedoraproject.org/rpms/intel-igc/c/96faceb1cb090c9902af4593cb8482b5fec8c109?branch=epel10.2
Log:
Add patch for LLVM15 segfault
---
diff --git a/0001-Fix-segfault-when-running-analysis-passes-after-SCCP.patch b/0001-Fix-segfault-when-running-analysis-passes-after-SCCP.patch
new file mode 100644
index 0000000..a65c808
--- /dev/null
+++ b/0001-Fix-segfault-when-running-analysis-passes-after-SCCP.patch
@@ -0,0 +1,30 @@
+From 4f7d1904031875fa5e8ed5d5b0d1e487bcc9b7b6 Mon Sep 17 00:00:00 2001
+From: Michal Paszkowski <michal.paszkowski@intel.com>
+Date: Fri, 5 Jun 2026 07:08:33 +0000
+Subject: [PATCH] Fix segfault when running analysis passes after SCCP
+
+SCCPLegacyPassWrapper runs new PM SCCPPass, which can change CFG and delete
+basic blocks. The wrapper incorrectly declared DominatorTreeWrapperPass as
+preserved without updating the legacy PM DomTree, leaving stale DomTree nodes
+that could reference freed BBs and crash later loop analyses.
+
+Fix by removing DomTree preservation in the SCCP wrapper so DT is invalidated
+and recomputed when required.
+---
+ IGC/WrapperLLVM/lib/llvmWrapper/Transforms/Scalar/SCCP.cpp | 1 -
+ 1 file changed, 1 deletion(-)
+
+diff --git a/IGC/WrapperLLVM/lib/llvmWrapper/Transforms/Scalar/SCCP.cpp b/IGC/WrapperLLVM/lib/llvmWrapper/Transforms/Scalar/SCCP.cpp
+index e1bf620380b3..b08854ef0de5 100644
+--- a/IGC/WrapperLLVM/lib/llvmWrapper/Transforms/Scalar/SCCP.cpp
++++ b/IGC/WrapperLLVM/lib/llvmWrapper/Transforms/Scalar/SCCP.cpp
+@@ -38,7 +38,6 @@ bool SCCPLegacyPassWrapper::runOnFunction(Function &F) {
+ void SCCPLegacyPassWrapper::getAnalysisUsage(AnalysisUsage &AU) const {
+ AU.addRequired<TargetLibraryInfoWrapperPass>();
+ AU.addPreserved<GlobalsAAWrapperPass>();
+- AU.addPreserved<DominatorTreeWrapperPass>();
+ }
+
+ char SCCPLegacyPassWrapper::ID = 0;
+--
+2.43.0
diff --git a/intel-igc.spec b/intel-igc.spec
index 68e1aea..6299ba5 100644
--- a/intel-igc.spec
+++ b/intel-igc.spec
@@ -48,6 +48,8 @@ Source5: https://github.com/KhronosGroup/SPIRV-Tools/archive/%{spirv_tools_commi
Source6: https://github.com/llvm/llvm-project/archive/refs/tags/llvmorg-%{llvm_ver}.tar.gz
%endif
+Patch0: 0001-Fix-segfault-when-running-analysis-passes-after-SCCP.patch
+
# This is just for Intel GPUs
ExclusiveArch: x86_64
reply other threads:[~2026-06-12 23:37 UTC|newest]
Thread overview: [no followups] expand[flat|nested] mbox.gz Atom feed
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=178130747248.1.11157231997824707851.rpms-intel-igc-96faceb1cb09@fedoraproject.org \
--to=pavel.androniychuk@intel.com \
--cc=git-commits@fedoraproject.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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox