Print :abs_gN: prefixes for ARM64 MOVZ/MOVK#20
Open
boppitybop wants to merge 3 commits intoGrammaTech:masterfrom
Open
Print :abs_gN: prefixes for ARM64 MOVZ/MOVK#20boppitybop wants to merge 3 commits intoGrammaTech:masterfrom
boppitybop wants to merge 3 commits intoGrammaTech:masterfrom
Conversation
added 2 commits
March 4, 2026 11:23
When a symbolic operand on a MOVZ/MOVK instruction carries G0-G3 attributes, print the appropriate :abs_gN: or :abs_gN_nc: prefix so the assembler emits the correct relocation. - G3 on MOVZ -> #:abs_g3:sym - G2 on MOVK -> #:abs_g2_nc:sym - G1 on MOVK -> #:abs_g1_nc:sym - G0 on MOVZ -> #:abs_g0:sym / G0 on MOVK -> #:abs_g0_nc:sym Companion PR: GrammaTech/ddisasm#90
boppitybop
force-pushed
the
fix/arm64-movz-movk-abs-prefix
branch
from
885d249 to
ecd2c09
Compare
junghee
suggested changes
Mar 6, 2026
Contributor
junghee
left a comment
junghee
left a comment
There was a problem hiding this comment.
@boppitybop Thank you for the contribution!
I've added some inline review comments -- please take a look and address them when you get a chance.
Also, could you add some test(s) in tests/arm64_syntax_test.py and include an entry in CHANGELOG.md?
Thanks again!
Comment on lines
+122
to
+150
| // Capstone aliases MOVZ as "mov", but the GNU assembler requires "movz" | ||
| // when using :abs_gN: relocation modifiers. Detect the MOVZ encoding from | ||
| // raw instruction bytes and use the canonical mnemonic, but only when the | ||
| // instruction has a symbolic operand with abs_g attributes. | ||
| if (opcode.empty() && inst.size == 4) { | ||
| uint32_t Enc = static_cast<uint32_t>(inst.bytes[0]) | ||
| | (static_cast<uint32_t>(inst.bytes[1]) << 8) | ||
| | (static_cast<uint32_t>(inst.bytes[2]) << 16) | ||
| | (static_cast<uint32_t>(inst.bytes[3]) << 24); | ||
| bool IsMovzEncoding = ((Enc >> 23) & 0x3F) == 0x25 | ||
| && ((Enc >> 29) & 0x3) == 0x2; | ||
| if (IsMovzEncoding) { | ||
| const gtirb::SymbolicExpression* Symex = | ||
| block.getByteInterval()->getSymbolicExpression( | ||
| ea - *block.getByteInterval()->getAddress()); | ||
| if (Symex != nullptr) { | ||
| const gtirb::SymAddrConst* Symaddr = | ||
| this->getSymbolicImmediate(Symex); | ||
| if (Symaddr != nullptr | ||
| && (Symaddr->Attributes.count(gtirb::SymAttribute::G0) | ||
| || Symaddr->Attributes.count(gtirb::SymAttribute::G1) | ||
| || Symaddr->Attributes.count(gtirb::SymAttribute::G2) | ||
| || Symaddr->Attributes.count(gtirb::SymAttribute::G3))) { | ||
| opcode = "movz"; | ||
| } | ||
| } | ||
| } | ||
| } | ||
|
|
Contributor
There was a problem hiding this comment.
Suggested change
| // Capstone aliases MOVZ as "mov", but the GNU assembler requires "movz" | |
| // when using :abs_gN: relocation modifiers. Detect the MOVZ encoding from | |
| // raw instruction bytes and use the canonical mnemonic, but only when the | |
| // instruction has a symbolic operand with abs_g attributes. | |
| if (opcode.empty() && inst.size == 4) { | |
| uint32_t Enc = static_cast<uint32_t>(inst.bytes[0]) | |
| | (static_cast<uint32_t>(inst.bytes[1]) << 8) | |
| | (static_cast<uint32_t>(inst.bytes[2]) << 16) | |
| | (static_cast<uint32_t>(inst.bytes[3]) << 24); | |
| bool IsMovzEncoding = ((Enc >> 23) & 0x3F) == 0x25 | |
| && ((Enc >> 29) & 0x3) == 0x2; | |
| if (IsMovzEncoding) { | |
| const gtirb::SymbolicExpression* Symex = | |
| block.getByteInterval()->getSymbolicExpression( | |
| ea - *block.getByteInterval()->getAddress()); | |
| if (Symex != nullptr) { | |
| const gtirb::SymAddrConst* Symaddr = | |
| this->getSymbolicImmediate(Symex); | |
| if (Symaddr != nullptr | |
| && (Symaddr->Attributes.count(gtirb::SymAttribute::G0) | |
| || Symaddr->Attributes.count(gtirb::SymAttribute::G1) | |
| || Symaddr->Attributes.count(gtirb::SymAttribute::G2) | |
| || Symaddr->Attributes.count(gtirb::SymAttribute::G3))) { | |
| opcode = "movz"; | |
| } | |
| } | |
| } | |
| } | |
| void Arm64PrettyPrinter::fixupInstruction(const gtirb::CodeBlock& block, cs_insn& inst) { | |
| ElfPrettyPrinter::fixupInstruction(block, inst); | |
| // Capstone aliases MOVZ as "mov", but the GNU assembler requires "movz" | |
| // when using :abs_gN: relocation modifiers. Detect the MOVZ encoding from | |
| // raw instruction bytes and use the canonical mnemonic, but only when the | |
| // instruction has a symbolic operand with abs_g attributes. | |
| if (inst.size == 4) { | |
| uint32_t Enc = static_cast<uint32_t>(inst.bytes[0]) | |
| | (static_cast<uint32_t>(inst.bytes[1]) << 8) | |
| | (static_cast<uint32_t>(inst.bytes[2]) << 16) | |
| | (static_cast<uint32_t>(inst.bytes[3]) << 24); | |
| bool IsMovzEncoding = ((Enc >> 23) & 0x3F) == 0x25 | |
| && ((Enc >> 29) & 0x3) == 0x2; | |
| if (IsMovzEncoding) { | |
| gtirb::Addr ea(inst.address); | |
| const gtirb::SymbolicExpression* Symex = | |
| block.getByteInterval()->getSymbolicExpression( | |
| ea - *block.getByteInterval()->getAddress()); | |
| if (Symex != nullptr) { | |
| const gtirb::SymAddrConst* Symaddr = | |
| this->getSymbolicImmediate(Symex); | |
| if (Symaddr != nullptr | |
| && (Symaddr->Attributes.count(gtirb::SymAttribute::G0) | |
| || Symaddr->Attributes.count(gtirb::SymAttribute::G1) | |
| || Symaddr->Attributes.count(gtirb::SymAttribute::G2) | |
| || Symaddr->Attributes.count(gtirb::SymAttribute::G3))) { | |
| memcpy(inst.mnemonic, "MOVZ", 5); | |
| } | |
| } | |
| } | |
| } | |
| } |
It would be better to utilize fixupInstruction in the base class.
Currently, it only takes cs_insn& inst, but we could extend it to also take the block:
- virtual void fixupInstruction(cs_insn& inst);
+ virtual void fixupInstruction(const gtirb::CodeBlock& block, cs_insn& inst);
For consistency, the overriding functions could then be updated accordingly:
-void AttPrettyPrinter::fixupInstruction(cs_insn& Insn) {
+void AttPrettyPrinter::fixupInstruction(const gtirb::CodeBlock& block,
+ cs_insn& Insn) {
+ PrettyPrinterBase::fixupInstruction(block, Insn);
+
Author
|
@junghee The latest commit should address all the comments, let me know if there is anything else when you get a chance to review again please, thanks! |
Contributor
Hi @boppitybop the changes are currently under internal review and should be merged soon. Thank you again for your contribution! |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Companion to GrammaTech/ddisasm#90.