https://github.com/kovdan01 created
https://github.com/llvm/llvm-project/pull/82364
Currently, for x86 and x86_64 triples, "+sse" and "+sse2" are appended to
`Features` vector of `TargetOptions` unconditionally. This vector is later
reset in `TargetInfo::CreateTargetInfo` and filled using info
https://github.com/kovdan01 ready_for_review
https://github.com/llvm/llvm-project/pull/82364
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/kovdan01 created
https://github.com/llvm/llvm-project/pull/82603
If `LLVM_TARGETS_TO_BUILD` does not contain `X86` and we try to run an x86
binary in lldb, we get a `nullptr` dereference in `LLVMDisasmInstruction(...)`.
We try to call `getDisAsm()` method on a `LLVMDisasmCon
kovdan01 wrote:
Please suggest ideas how this could be tested. It looks like that
UnwindAssembly unittests is the right place for it - but UnwindAssemblyx86Tests
cmake target is not built without X86 in `LLVM_TARGETS_TO_BUILD`, which is the
pre-condition of the issue this PR is fixing.
https:
https://github.com/kovdan01 ready_for_review
https://github.com/llvm/llvm-project/pull/82603
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
kovdan01 wrote:
> tbh I have no problems with the patch, but I think it's fixing something that
> I think should be reconsidered altogether, I'm interested to hear more about
> what the use case looks like that led to this being a problem.
@jasonmolenda The use case is very simple, describing
kovdan01 wrote:
> Would this change be observable by a test?
@adrian-prantl Theoretically, it should be: in
`ClangExpressionParser::ClangExpressionParser`, we try to hardcode sse and sse2
for both x86 and x86_64, while in `X86TargetInfo::initFeatureMap`, sse2
(implying sse) is only hardcoded
kovdan01 wrote:
> Can this code be hit when using an x86 core file?
Thanks for suggestion! I'll try that and notify here if such approach succeeded.
https://github.com/llvm/llvm-project/pull/82603
___
lldb-commits mailing list
lldb-commits@lists.llvm.
https://github.com/kovdan01 closed
https://github.com/llvm/llvm-project/pull/82364
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/kovdan01 created
https://github.com/llvm/llvm-project/pull/84387
Depends on #84384
This adds support for `DW_TAG_LLVM_ptrauth_type` entries corresponding
to explicitly signed types (e.g. free function pointers) in lldb user
expressions. Applies PR https://github.com/apple/ll
https://github.com/kovdan01 updated
https://github.com/llvm/llvm-project/pull/84387
>From 728f5644aebfafd2114e7e47a9b83ef057423997 Mon Sep 17 00:00:00 2001
From: Jonas Devlieghere
Date: Tue, 20 Feb 2024 10:57:54 -0800
Subject: [PATCH 1/4] Upstream ptrauth changes to DWARFASTParserClang
---
..
https://github.com/kovdan01 ready_for_review
https://github.com/llvm/llvm-project/pull/84387
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
kovdan01 wrote:
> Can this code be hit when using an x86 core file? Then you could write a
> shell test thatis `UNSUPPORTED: x86-registered-target` (whatever the proper
> syntax is) and assert that it does not crash.
@DavidSpickett The issue is present when loading an x86 core files via `lldb
@@ -0,0 +1,10 @@
+# UNSUPPORTED: x86
kovdan01 wrote:
Yes, I've checked that with `X86` in `LLVM_TARGETS_TO_BUILD`, the test becomes
unsupported. So, it looks like it's currently OK
https://github.com/llvm/llvm-project/pull/82603
kovdan01 wrote:
> As this plugin seems related to backtrace perhaps the test should at least
> run `bt` and check for the first line of the output.
I had similar thoughts, but I'm not sure if placing the test in
`lldb/test/Shell/Commands` as `command-backtrace-missing-x86.test` is a nice
idea
kovdan01 wrote:
@jasonmolenda Just in case you've missed - I've provided the use case
description leading to the issue (as you requested) above
https://github.com/llvm/llvm-project/pull/82603#issuecomment-1970019956. Would
be glad if you let me know if it gives you enough info & if this partic
kovdan01 wrote:
A kind reminder regarding the PR - would be glad to see feedback from everyone
interested.
https://github.com/llvm/llvm-project/pull/84387
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman
https://github.com/kovdan01 updated
https://github.com/llvm/llvm-project/pull/84387
>From 728f5644aebfafd2114e7e47a9b83ef057423997 Mon Sep 17 00:00:00 2001
From: Jonas Devlieghere
Date: Tue, 20 Feb 2024 10:57:54 -0800
Subject: [PATCH 01/10] Upstream ptrauth changes to DWARFASTParserClang
---
https://github.com/kovdan01 updated
https://github.com/llvm/llvm-project/pull/84387
>From 728f5644aebfafd2114e7e47a9b83ef057423997 Mon Sep 17 00:00:00 2001
From: Jonas Devlieghere
Date: Tue, 20 Feb 2024 10:57:54 -0800
Subject: [PATCH 1/6] Upstream ptrauth changes to DWARFASTParserClang
---
..
@@ -664,6 +685,17 @@ CompilerType CompilerType::GetPointerType() const {
return CompilerType();
}
+CompilerType
+CompilerType::AddPtrAuthModifier(unsigned key, bool isAddressDiscriminated,
kovdan01 wrote:
> Also, whose this user of this API?
I've updated t
@@ -216,6 +216,16 @@ class TypeSystem : public PluginInterface,
virtual uint32_t GetPointerByteSize() = 0;
+ // TODO: are we allowed to insert virtual functions in the middle of the
class
+ // interface and break ABI?
kovdan01 wrote:
Thanks, deleted TOD
@@ -676,6 +677,62 @@ DWARFASTParserClang::ParseTypeModifier(const SymbolContext
&sc,
case DW_TAG_volatile_type:
encoding_data_type = Type::eEncodingIsVolatileUID;
break;
+ case DW_TAG_LLVM_ptrauth_type: {
+DWARFDIE ptr_die = die.GetReferencedDIE(DW_AT_type);
---
kovdan01 wrote:
@jasonmolenda Would be glad to see your feedback - see answers to your previous
comments above
https://github.com/llvm/llvm-project/pull/82603
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mai
kovdan01 wrote:
@jasonmolenda A kind reminder regarding the PR - see answers to your previous
comments above
https://github.com/llvm/llvm-project/pull/82603
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailm
kovdan01 wrote:
Thanks @jasonmolenda for your feedback and suggestion! See
f96989dd1f832284b74d07d1e457a15a0b16c199 - I've deleted the test with corefile
and added the test you've mentioned. Basically, I've just left the most simple
test from "normal" `Testx86AssemblyInspectionEngine` and chec
https://github.com/kovdan01 closed
https://github.com/llvm/llvm-project/pull/82603
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/kovdan01 updated
https://github.com/llvm/llvm-project/pull/84387
>From 728f5644aebfafd2114e7e47a9b83ef057423997 Mon Sep 17 00:00:00 2001
From: Jonas Devlieghere
Date: Tue, 20 Feb 2024 10:57:54 -0800
Subject: [PATCH 1/6] Upstream ptrauth changes to DWARFASTParserClang
---
..
https://github.com/kovdan01 edited
https://github.com/llvm/llvm-project/pull/84387
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/kovdan01 updated
https://github.com/llvm/llvm-project/pull/84387
>From 728f5644aebfafd2114e7e47a9b83ef057423997 Mon Sep 17 00:00:00 2001
From: Jonas Devlieghere
Date: Tue, 20 Feb 2024 10:57:54 -0800
Subject: [PATCH 1/6] Upstream ptrauth changes to DWARFASTParserClang
---
..
kovdan01 wrote:
Merging this previously approved PR since prerequisites are now met
https://github.com/llvm/llvm-project/pull/84387
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
https://github.com/kovdan01 closed
https://github.com/llvm/llvm-project/pull/84387
___
lldb-commits mailing list
lldb-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/lldb-commits
31 matches
Mail list logo
